In the autumn of 1993, MG Refining and Marketing — the United States energy subsidiary of the German industrial conglomerate Metallgesellschaft AG — held a position that, on paper, was a textbook hedge. The firm had sold long-term forward contracts to deliver heating oil, gasoline, and diesel to commercial customers at fixed prices, in some cases ten years out. To offset the resulting price exposure, MGRM had built a roughly equal notional hedge using near-dated NYMEX energy futures and short-dated over-the-counter swaps, rolled forward month by month — a stack-and-roll hedge.

The economics looked clean in spot-price space. If oil rose, the long-term supply contracts became more onerous, but the hedge appreciated. If oil fell, the contracts gained value, the hedge lost. Against directional spot moves, the hedge was substantially complete.

Through the second half of 1993, oil prices fell sharply. WTI crude declined from around $20 a barrel in early June to about $14 by mid-December. The hedge position, marked to market daily, generated cash margin calls on a scale that consumed MGRM’s available funding. The supply contracts also gained economic value under the spot-price logic of the hedge, but those gains were long-dated, illiquid, and contested in valuation. Margin calls demanded cash today; offsetting gains existed only on a discounted-cash-flow schedule running out to 2003.

By December 1993, the supervisory board of the German parent, alarmed at margin calls in the hundreds of millions of dollars, dismissed executive chairman Heinz Schimmelbusch and appointed a new management team that liquidated the futures positions and unwound the customer contracts. The realised loss, on the auditors’ calculation, was approximately $1.3 billion. To the extent those gains existed, they were not monetised on a timetable that funded the margin calls, and much was lost when the programme was terminated.

The puzzle is not whether the hedge was right in theory. Subsequent academic work continues to debate that, with serious arguments on both sides. The puzzle is structural. A position designed to match one projection of price exposure failed catastrophically when other control-relevant dimensions — basis, maturity mismatch, and the timing of cash flows under funding stress — became binding.

What kind of problem is this?

When a control system maps fewer dimensions than the system it is meant to control, when the regulator’s response set covers some kinds of disturbance and is silent on others, when the disturbance the regulator cannot recognise is the one that arrives — that is a specific shape of problem. And the discipline that formalised it, in the middle of the twentieth century, is cybernetics.

So let’s borrow.

This is the multi-model move: recognise the shape of a problem, find the discipline that has thought rigorously about that shape, and import its frameworks deliberately rather than reinventing them from scratch.

Cybernetics was developed in the 1940s and 1950s by Norbert Wiener, W. Ross Ashby, and others, working initially on biological control, electrical engineering, and the design of self-regulating systems — thermostats, gun-laying servos, neural feedback. Markets do not satisfy its assumptions cleanly: participants are reflexive and adaptive, regulation reshapes what is regulated, and disturbances arrive in forms no designer can enumerate in advance. But the structural logic — that any regulator must match the variety of the system it confronts, and must internally embody a model of that system on the dimensions that matter — transfers.

Where Variety Was Missing

W. Ross Ashby, a British psychiatrist who became one of the founders of cybernetics, published An Introduction to Cybernetics in 1956. Its central result, the Law of Requisite Variety, is one of the most compact statements in the formal study of control. Ashby’s slogan was: only variety can destroy variety.

The technical content is precise. Variety is the number of distinguishable states a system can occupy. A coin has variety two; a six-sided die has variety six; a portfolio with twenty independent risk factors, each potentially in any of several regimes, has variety in the millions. In its information-theoretic form, Ashby’s law sets a floor on residual variety: the variety appearing at the essential variable cannot be reduced below the variety of the disturbance set minus the effective variety of the regulator, unless the structure of the system itself attenuates it.

The implication is structural rather than tactical. There is a minimum effective complexity below which the specified degree of control is impossible. No amount of speed, discipline, or effort substitutes for variety the regulator does not possess.

Apply this to MGRM. In spot-price space, the hedge had substantial variety. But matching variety was insufficient across basis, maturity, and funding space. A fall in spot prices generated immediate cash losses on the stacked short-dated hedge, while the offsetting gains emerged only gradually, depended on customer behaviour, and were not pledgeable as collateral. Contango was a term-structure disturbance that produced roll losses and funding stress. The regulator suppressed one channel of disturbance while leaving others open; in Ashby’s terms, residual variety had a route to the essential variable when the liquidity disturbance arrived.

The Hidden Map

In 1970, Ashby and Roger Conant published a paper with one of the more striking titles in the cybernetic literature: “Every Good Regulator of a System Must Be a Model of That System.” The result extends the law of requisite variety into a structural theorem about the regulator’s internal organisation.

The Conant-Ashby theorem, stated informally, holds that the simplest optimal regulator must contain a model of the system, in the sense that the regulator’s events map the relevant events of the system. A regulator that does not encode such a mapping, on the dimensions where regulation is required, cannot be the simplest optimal regulator there.

The consequence is practical. A risk control system that does not model funding dynamics cannot regulate funding risk optimally, however sophisticated its handling of price risk. A governance protocol that does not model the difference between mark-to-market and economic loss cannot regulate the response to drawdowns optimally, however disciplined the decision-makers.

MGRM provides two instances. The trading desk’s hedge was a model of one projection of price dynamics, and a precise one. It was not a sufficiently rich model of basis, maturity mismatch, and funding dynamics. In Conant-Ashby terms, the regulator did not map the system on the dimensions that later became decisive.

The Frankfurt supervisory board, by late 1993, faced the position through a frame in which mark-to-market losses were salient and the long-dated supply contracts’ offsetting value less so. Whether a richer frame would have led to a different choice is contested — Culp and Miller argued yes, Mello and Parsons no, on the grounds funding stress made continuation infeasible regardless. Either way, the board needed a usable model of the joint state — economic value, funding capacity, governance tolerance — and the regulator on the second floor had to map a different system from the regulator on the trading floor.

Where Compression Costs

The temptation to reduce variety is structural in capital allocation. Variety is expensive to monitor, govern, and explain. Compression makes oversight tractable, simplifies committee discussion, and lets a complex system be presented to non-specialist principals. In each of three places, the price of compression is paid in residual variety reaching the goal.

The first is risk measurement that distils a multi-dimensional system into a single number. Value-at-risk, expected shortfall, single-figure stress-test outputs — each is a useful summary and a poor standalone regulator. They compress a high-dimensional state into one measurement channel. The danger is not the scalar itself but the response rule attached to it. If the rule does not preserve the distinctions that matter for action — funding liquidity, basis, crowdedness, counterparty fragility, liquidation horizon — the missing variety reappears at the essential variable.

The second is governance escalation paths that map one dimension. A protocol that says “intervene if drawdown exceeds X” maps drawdown. The dimensions it omits — counterparty stress, headline risk, succession dynamics, the difference between permanent capital impairment and recoverable mark-to-market loss — are precisely where senior intervention shapes outcomes. The protocol’s variety is fixed; the situation’s variety is not.

The third is rule-based portfolio construction that selects a fixed response to a variable system. Static allocation grids, fixed-period rebalancing schedules, and tilt strategies with constant weights all have determined response variety. Markets do not. When the regime occupies a state the rule does not distinguish, the rule responds with whatever it has, which may be wrong in the dimension that has changed.

None of these compressions are wrong as summaries or as starting points. They become wrong when they are mistaken for regulators.

Two earlier posts intersect this framework. The Path The Maths Misses showed that a calculation correct on the ensemble average can be ruinous on the realised time-path. Variety mismatch is the cybernetic complement: a regulator whose response set spans the average-path disturbance but not the realised-path disturbance is undermatched in exactly that sense.

The Model That Fights Back described why a system under pressure intervenes harder rather than revising its model — the precision trap, the complexity cost of moving prior beliefs. Coupled with requisite variety, the trap closes mechanically: an under-varietied regulator cannot afford to acknowledge missing dimensions, because acknowledgement requires rebuilding rather than running harder. Conant-Ashby suggests why a richer model would have been required for an optimal response. The free energy framework suggests why such a model is hard to build under pressure.

The Capacity Question

When a control system is failing, the first question is not whether it is being applied with sufficient discipline but whether it has the variety to recognise and respond to the disturbance arriving. If not, more discipline cannot rescue it; only more effective variety — or a redesign that attenuates the disturbance before it reaches the controlled variable — can. Running an under-varietied regulator harder produces faster failure, not eventual success.

The diagnosis applies wherever a control system maintains something against disturbance. In organisations, a control system that maps performance against budget but not capability accumulation will under-regulate the dimension that matters most for long-run survival. In professional practice, a quality regime that maps procedure but not judgement leaves residual variety to whatever individuals improvise. In health, a regimen tracking measurable biomarkers but not subjective state under-regulates the channel through which much of the relevant disturbance arrives.

The discipline is in asking, not whether the regulator is being run well, but whether it has the structural variety to recognise and respond to what it is being asked to control?

References & Further Reading

1. An Introduction to Cybernetics: W. Ross Ashby

2. Design for a Brain: The Origin of Adaptive Behaviour: W. Ross Ashby

3. Every Good Regulator of a System Must Be a Model of That System: Roger C. Conant & W. Ross Ashby

4. Metallgesellschaft and the Economics of Synthetic Storage: Christopher L. Culp & Merton H. Miller

5. Maturity Structure of a Hedge Matters: Lessons from the Metallgesellschaft Debacle: Antonio S. Mello & John E. Parsons

6. Metallgesellschaft: A Prudent Hedger Ruined, or a Wildcatter on NYMEX?: Stephen Craig Pirrong

7. Derivatives Debacles: Case Studies of Large Losses in Derivatives Markets: Anatoli Kuprianov

8. The Cybernetic Brain: Sketches of Another Future: Andrew Pickering

On the morning of 16 September 1992, sterling was pinned to its lower bound inside the European Exchange Rate Mechanism. The framework had been clear for nearly two years. Sterling would hold within a band against the Deutsche Mark. The Bank of England would defend the band. Domestic monetary policy would be subordinated to it. In return, Britain would import the Bundesbank’s anti-inflationary credibility.

By that morning, every part of the framework was being tested simultaneously. The Bundesbank had raised rates to absorb the fiscal shock of German reunification. Britain was in recession, with rising unemployment and falling property prices. Domestic conditions called for lower rates. The framework called for keeping UK rates high enough to defend the parity. The market noticed the gap.

At around 11:00 the government announced an increase in the base rate from 10 to 12 per cent. Sterling continued to fall. An emergency announcement promised a further increase to 15 per cent the following morning. Sterling continued to fall. By 19:40, the Chancellor of the Exchequer announced that the United Kingdom was suspending its membership of the ERM. The 15 per cent rate would never take effect. Treasury papers declassified in 2005 put the estimated loss at around £3.3 billion, measured at February 1994.

The framework had been internally coherent. It bought one desirable property — anti-inflationary credibility — at the price of another, monetary autonomy. As long as those two were not in conflict, the price was unobservable. The day they came into conflict, it was paid in a single afternoon.

What kind of problem is this?

The ERM was not a bad framework that turned out to be wrong. It was a framework with a particular structure: a set of desirable properties that could not all be jointly held. Free capital movement, a fixed exchange rate, and independent monetary policy form what Robert Mundell formalised in 1963 as the impossible trinity. Any two are achievable; all three are not. Britain by 1992 had committed to the first two and was discovering that the third was not available.

When a system of desirable properties cannot all be held simultaneously, when committing to one set forces a choice about which to give up, when the choice itself is the cost of coherence — that is a specific shape of problem. And the discipline that has thought most rigorously about systems of properties that cannot all be satisfied is the foundations of quantum mechanics.

So let’s borrow.

This is the multi-model move: recognise the shape of a problem, find the discipline that has thought rigorously about that shape, and import its frameworks deliberately rather than reinventing them from scratch.

The foundations of quantum mechanics work with entangled particles and photon pairs, physical systems whose correlations can be measured in the laboratory. Markets are messier: there is no equivalent apparatus, participants adapt, and committing to a framework shapes the portfolio. But the structural logic transfers. When a set of desirable properties is mutually inconsistent, every coherent position must surrender one of them, and the choice of which to surrender is the price of taking a position.

Three Truths, At Most Two

In 1964, the Northern Irish physicist John Stewart Bell published “On the Einstein Podolsky Rosen Paradox”. It addressed whether quantum mechanics is a complete description of reality, or whether hidden variables would restore a more classical picture in which particles have definite properties at all times.

Bell took three assumptions that any sensible classical theory ought to satisfy. The first was locality: nothing influences anything else faster than light. The second was definite outcomes: measurements yield single, classical values, and those values reflect properties the system possessed prior to being measured. The third was measurement independence: experimenters can choose what to measure freely, without their choices being correlated with the hidden state of the system.

From these assumptions, Bell derived an inequality that set a ceiling on how strongly distant measurements of entangled particles could correlate. Quantum mechanics predicts that this ceiling can be exceeded. The two predictions disagree by an amount testable in the laboratory.

In 1982, Alain Aspect’s group in Paris ran a landmark version of the experiment. Quantum mechanics won. In 2015, three independent groups closed the main experimental loopholes. Quantum mechanics won again. In 2022, the Nobel Prize in Physics was awarded to Aspect, John Clauser, and Anton Zeilinger for this work.

The implication is uncomfortable. Those three assumptions cannot all be true together. Locality, definite outcomes, or measurement independence: one must go. The empirical result establishes that. It does not tell us which.

Four Ways To Pay

The leading responses to Bell’s theorem either reinterpret the standard quantum formalism, as Many Worlds, Bohmian mechanics, and Copenhagen-type views do, or modify its assumptions, as superdeterminism does. What separates them is which of Bell’s assumptions each is prepared to surrender.

Many Worlds, originated by Hugh Everett in 1957 and developed by Bryce DeWitt and David Deutsch, keeps dynamical locality, measurement independence, and the realism of the underlying wavefunction. The price is the single-outcome component of definite outcomes: there is no unique result, but a branching structure in which every quantum-allowed outcome is realised in some branch.

The pilot-wave or Bohmian interpretation, developed by Louis de Broglie and David Bohm, keeps definite outcomes — particles have positions at all times — and measurement independence. The price is locality at the hidden-variable level: the guiding wavefunction ties distant particles together, though not in a way that permits faster-than-light signalling. Lee Smolin notes a further asymmetry: the wave influences the particle but the particle does not influence the wave. That asymmetry is part of the price.

The Copenhagen interpretation and its modern descendants keep operational locality and measurement independence. The price is the other component of definite outcomes: classical values cannot be assigned to observables outside a measurement context. The wavefunction describes potentialities; values come into being with the act of measurement. The result is a strange relationship between observer and observed that has bothered physicists for a century.

Superdeterminism, advocated by Sabine Hossenfelder and others, keeps locality and definite outcomes by giving up measurement independence. Experimenters’ choices are correlated with the hidden state of the system being measured. Critics argue that this dissolves the basis of experimental reasoning; proponents argue that the price, properly understood, is bearable.

Each of these is a coherent response to Bell’s constraint. The first three are usually presented as empirically equivalent interpretations of the standard quantum formalism; superdeterminism is more radical, modifying assumptions rather than reinterpreting them. Bell’s theorem proves that something must be surrendered. The position ruled out is the one that tries to keep all the incompatible commitments at once.

What Every Framework Surrenders

Investing under uncertainty has an analogous structural shape, even though the mathematics is different. The desirable properties that frameworks for capital allocation try to hold are abundant: high expected return, low volatility, resilience to drawdown, liquidity, regime independence, low cost, robustness to behavioural error. No portfolio holds all of them at once. The framework is a choice about which to surrender.

Naïve mean-variance optimisation gives up robustness to input error and regime change. The framework assumes returns are stationary and covariances estimable from the past. When the distribution that generated yesterday’s data is not the distribution generating tomorrow’s, the optimisation computes the wrong answer with great precision.

Risk parity gives up resilience to a world in which its diversifiers stop diversifying. The framework equalises risk contributions across asset classes by levering the lower-volatility ones, premised on diversifying behaviour that held through the 2010s. In 2022, when bonds and equities fell in concert, the leverage that had been the framework’s elegance became its liability.

Trend-following gives up the turning point. The framework profits when prices continue in their current direction and loses when they reverse. The price is paid at every reversal, when the strategy is committed to a direction the market has just abandoned.

Concentrated value gives up the resilience that comes from diversification. The framework trusts that conviction in a small number of positions outweighs the protection of being wrong about any one. When the theses fail together, the concentration carries the correlation it appeared to avoid.

Volatility selling gives up convexity. The framework collects steady option premium against the assumption that realised volatility will fall short of implied. When the assumption holds, the income is reliable; when it fails, the losses can arrive in concentrated bursts that outweigh years of accumulated premium.

Carry strategies give up resilience in risk-off regimes. The framework borrows cheaply in one currency or maturity and lends at higher yields in another, harvesting the spread. The unwind, when it arrives, can compress months of accumulated carry into days of forced selling.

Every framework names what it surrenders, even if adherents find it convenient not to look sometimes. The most expensive framework in any room is the one whose proponents claim it costs nothing. Frameworks whose costs are honestly identified can be hedged, complemented, or held with humility. Frameworks whose costs are denied compound until the world demands payment, as the ERM compounded from October 1990 until September 1992, and as LTCM’s framework compounded for four years until a New York Fed boardroom filled, in September 1998, with representatives of the institutions called upon to settle the cost.

Knowing What You Owe

When physicists weigh these interpretations against one another, the experiments cannot adjudicate among them. The debate turns on what each interpretation keeps, what it surrenders, and whether the price is acceptable. They are not free to escape Bell’s price. They are free only to choose how to pay it.

The same discipline is available to those who allocate capital. No framework escapes the price. The unknowns are too dense, the data too sparse, the regimes too non-stationary for any single framework to be both complete and free. The framework is its trade-off, not its content.

This lesson extends beyond finance. In engineering, every design surrenders something: weight reduction can cost fatigue resistance, redundancy costs efficiency. In clinical medicine, every treatment surrenders something: aggressive intervention costs tolerability, sensitive screening costs specificity. The professional who claims a design or regimen with no downside has not escaped the trade-off; they have hidden it.

What follows is not relativism. Some frameworks are better than others for a given problem. Bell’s theorem itself rules out positions that try to keep everything. But within the space of internally coherent positions, the choice is between which truth you find least painful to give up. Pretending no truth is being given up is the only position provably wrong.

The discipline is in asking: what is the price my framework is paying, and am I paying it knowingly?

References & Further Reading

1. Speakable and Unspeakable in Quantum Mechanics: John Stewart Bell

2. Einstein’s Unfinished Revolution: Lee Smolin

3. The Quantum Challenge: Modern Research on the Foundations of Quantum Mechanics: George Greenstein and Arthur Zajonc

4. International Economics: Robert Mundell

5. Manias, Panics, and Crashes: A History of Financial Crises: Charles P. Kindleberger and Robert Z. Aliber

In late October 1988, the board of RJR Nabisco received what was, on the surface, a straightforward decision. The company’s chief executive, Ross Johnson, had proposed a management-led leveraged buyout that would take the company private. Within weeks, Kohlberg Kravis Roberts had submitted a competing bid. By the time the board announced KKR’s victory on 30 November, the contest had drawn international attention and produced one of the largest leveraged buyouts in history. The eventual price, around $25 billion, would stand as a record for years.

The conventional account of the auction focuses on the price competition. The bids escalated. The financing structures grew more complex. What received less attention was the structural fact at the centre of the case. The auction was managed by a special committee of the RJR board, chaired by Charles Hugel, which the directors had constituted for that purpose. The committee defined the criteria. The committee evaluated the bids against those criteria. The committee supplied the recommendation around which the full board’s decision formed. Whatever each bidder offered in headline price, the operative question was always how the bid would land with the committee under the criteria the committee itself had set.

By 30 November, when the board announced its decision, the headline numbers had converged closely enough that the price differential was not the decisive factor. KKR had structured its bid in a way the committee judged superior — employee protections, a plan to keep more of the company intact, more credible financing, a stronger securities package for existing shareholders, and a cleaner governance path after the transaction. The management group, despite having been the architects of the original transaction and despite a competitive headline price, did not prevail because the committee’s criteria were not reducible to price alone.

The puzzle is straightforward. The headline weights — the bid prices — were not the operative quantity. The operative quantity was the committee’s structural position: no auction outcome was likely to survive the process without the committee’s affirmative recommendation, governed by criteria broader than price. Bidders who treated the auction as a price contest misread the structure of the decision.

What kind of problem is this?

When the visible quantity — claim size, vote weight, headline allocation — diverges systematically from the operative quantity, which is how often an actor is the one whose agreement an outcome cannot do without, that is a specific shape of problem. And the discipline that has formalised the conditions under which power, in any voting or coalitional system, can be measured rigorously is cooperative game theory.

So let’s borrow.

This is the multi-model move: recognise the shape of a problem, find the discipline that has thought rigorously about that shape, and import its frameworks deliberately rather than reinventing them from scratch.

Cooperative game theory was developed by John von Neumann and Oskar Morgenstern in the 1940s and refined by Lloyd Shapley in the 1950s, built for abstract games with well-defined rules, payoffs, and players. Restructurings and corporate governance do not satisfy its assumptions cleanly: actors do not always have well-defined preferences, value cannot always be freely transferred between parties, the rules themselves are sometimes negotiable. But the structural logic — the habit of asking who is pivotal rather than who is largest, and the recognition that the gap between the two is where misjudgment of power lives — transfers cleanly. The vocabulary for marginal contribution to coalitions and the axiomatic basis for power allocation transfers.

Where Power Sits

A player is pivotal, in the technical sense, when their joining is the move that converts a losing coalition into a winning one — when their addition is the threshold-crosser. Pivotality is the formal property of how often an actor occupies that position across the structure of possible coalitions.

Lloyd Shapley’s result, circulated by RAND in 1952 and published canonically in 1953, is one of the cleanest in mathematical economics. He proved that there is a unique way to allocate the total value created by a coalition across its members that satisfies four axioms.

The first is efficiency: the total value is fully distributed across the players, with nothing left over. The second is symmetry: two players who contribute identically to every coalition they could join receive equal allocations. The third is linearity: when two coalition games are combined, the allocation in the combined game is the sum of the allocations in the components. The fourth is the dummy property: a player who contributes nothing to any coalition receives nothing.

Any allocation rule satisfying all four axioms is the same rule. Shapley showed it has a specific form: each player’s allocation is their average marginal contribution across all possible orderings in which the coalition could form. In the special case of a simple voting game, this becomes a direct measure of pivotality. A player who never tips the coalition has Shapley value zero, regardless of headline weight.

The application of Shapley’s framework to formal voting bodies — committees, legislatures, weighted majority games — was developed by Shapley and Martin Shubik in 1954; the resulting Shapley-Shubik power index is the canonical measure of pivotality in voting systems. A short worked example makes the point: a three-member board with voting weights of 49, 49, and 2, and a simple-majority threshold. Any minimal winning coalition requires exactly two members. Run through the orderings: the 2 per cent member tips the coalition as often as either 49 per cent member. Shapley value: one-third each. Headline weight, in this case, misleads completely.

The Banzhaf index, proposed by John Banzhaf in 1965 in a paper analysing the Nassau County Board of Supervisors, supplies a closely related measure — the probability that a player is pivotal in a randomly drawn coalition rather than the marginal contribution averaged across orderings. It often gives different numerical answers but the same structural insight.

The point is not the specific number. A rigorous treatment of power in any voting or coalitional system can produce an allocation that bears little resemblance to the headline weights. RJR is not a literal weighted-voting game — the committee was an institutional gatekeeper, not a 2 per cent voter — but the structural lesson transfers: the actor whose assent converts a proposal into an outcome can hold power far in excess of any headline economic exposure.

Where Headline Weight Misleads

The pattern recurs wherever multiple actors with formal voting weights or structural positions must reach a threshold to act. In investing, three places matter, and each is sustained by a specific mechanism that produces the gap between pivotality and headline weight.

The first is where the threshold is set, not crossed — board committees and structured auction processes. RJR is the leading example. When a board constitutes a special committee with effective authority over the sale process, the committee becomes pivotal because it owns the definition of the threshold itself. Its criteria are the operative constraint. The economic exposures of the parties being weighed are not. Even within ordinary voting games, the threshold dominates the weights: with holdings of 45, 45, and 10 and a simple-majority quota, the 10 per cent holder has the same formal pivotality as either 45 per cent holder. Raise the quota to 90 per cent and the same holder becomes irrelevant, because the two larger blocs together meet the threshold while 45 plus 10 falls short. The effect is mechanical but not monotone, and it has to be computed. Bidders and counterparties who model the situation as a contest of headline weights miss the operative quantity entirely.

The second is where the larger blocs disagree — governance and proxy contests. When two large holders align and together clear the threshold, a smaller third holder is practically a dummy. When they oppose, and neither side can reach the threshold alone, the same small holder can become decisive. Pivotality is conditional on the alignment structure, not on the weights alone. This is where the classical indices reach the limit of their usefulness: Shapley-Shubik and Banzhaf are a priori measures that abstract from actual preferences and treat coalitions according to specified symmetry assumptions. In institutional analysis, the relevant pivotality is preference-conditional. The formal weights are the starting point; the alignment model determines which coalitions are realistic.

The third is where structural position outruns claim size — gatekeeping intermediaries with statutory or contractual rights. Trustees in bond indentures, agents in syndicated loan agreements, fiscal agents in some kinds of transaction — each can hold influence not reducible to economic exposure. The mechanism that creates this divergence is legal or contractual; the mechanism that causes others to misprice it is informational. Counterparties who fail to recognise the gatekeeper’s structural position misallocate effort, competing on the visible quantity rather than positioning for the operative one. In the RJR case, this is what the bidders treating the auction as a price contest got wrong; in restructurings, it is what creditors focused on claim seniority routinely miss.

This connects to two earlier posts. What Others Think Others Think showed how higher-order beliefs — what each actor thinks others think, recursively — can dominate first-order analysis in coordination games; the pivotality structure of a coalition is the cooperative-game-theory analogue, where pivotality depends on what each actor believes about the alignments of the others. The Network You Cannot See showed that systemic risk lives in the edges between participants. The Shapley framework supplies a complementary insight: power lives in the pivotality structure between participants, and headline weights tell you about nodes, not edges.

The Diagnostic

The three places map to three diagnostic questions. Before judging power in any coalition situation, work through them in order.

What is the threshold structure? Simple majority, supermajority, unanimity? Different thresholds produce different pivotality distributions over the same voting weights, and the mapping is not intuitive without explicit analysis.

What is the correlation of preferences among the large actors? Pivotality is conditional on this, and the small holder’s position can shift from decisive to irrelevant as alignments shift.

Where is pivotality sitting given those two answers? The answer often diverges sharply from headline weight, and the divergence is where institutional intuition needs to be retrained.

The framing applies wherever multiple actors with formal voting weights must reach a threshold to act. In public policy, legislative coalitions exhibit the same pattern: small parties hold disproportionate power under supermajority rules when the larger parties divide. In standards bodies, technical specification authors with formally equal voting rights have unequal pivotality given correlation patterns among the larger participants. In multi-employer pension plans and joint ventures, the headline structure of voting weights and the operative structure of pivotality routinely diverge.

The discipline is in asking: not who owns the largest share, but who would the coalition need to win, and how does that change when the threshold changes?

References & Further Reading

1. A Value for n-Person Games: Lloyd Shapley

2. Theory of Games and Economic Behavior: John von Neumann & Oskar Morgenstern

3. The Model Thinker: Scott E. Page

4. Weighted Voting Doesn’t Work: A Mathematical Analysis: John F. Banzhaf III

5. Barbarians at the Gate: The Fall of RJR Nabisco: Bryan Burrough & John Helyar

6. A Method for Evaluating the Distribution of Power in a Committee System: Lloyd Shapley & Martin Shubik

7. RJR Nabisco: A Case Study of a Complex Leveraged Buyout: Allen Michel & Israel Shaked

In February 1994, the Federal Reserve raised its federal funds target for the first time in five years. The move caught markets off guard — leveraged bond portfolios suffered immediate losses. But most survived. Then the Fed raised again in March. And again in April. Over the course of the year, rates rose from 3 per cent to 5.5 per cent across six increments.

The first shock was sharp but absorbable. What followed was worse. Each subsequent rate rise eroded holdings, consumed capital, and tightened leverage ratios in structures already weakened by February’s blow. The portfolio that survived the initial surprise was slightly weaker entering March. The one that survived March was weaker still entering April. By the time the fifth and sixth increases arrived, funds that had comfortably absorbed any individual move were running on depleted reserves and unable to withstand shocks they would have shrugged off nine months earlier. Several prominent funds collapsed; Orange County declared what was then the largest municipal bankruptcy in American history.

The risk models had been asked: can this portfolio survive a rate rise of this magnitude? They had answered correctly — for any single move, yes. The question they had not been asked was: what happens to the portfolio’s capacity to absorb shocks after it has already absorbed several in succession?

What kind of problem is this?

When a system appears robust under any individual stress but fails after repeated exposure; when failure originates not at the point of greatest average load but at features that concentrate stress locally; when damage accumulates invisibly and then propagates suddenly — that is a specific shape of problem.

And the discipline that has spent a century studying how flaws concentrate stress is fracture mechanics.

So let’s borrow.

This is the multi-model move: recognise the shape of a problem, find the discipline that has thought rigorously about that shape, and import its frameworks deliberately rather than reinventing them from scratch.

Fracture mechanics was built for metals, ceramics, and composites — materials with well-characterised properties under controlled loading. Markets are messier: participants adapt, structures evolve, and measuring risk can change the system being measured. But the structural logic — the habit of asking where stress concentrates and how damage accumulates under repeated loading — transfers.

Strength and Its Flaws

The discipline begins with a puzzle that troubled engineers for decades: materials fail at loads far below their theoretical strength. In 1921, the British engineer A.A. Griffith resolved this by showing that fracture is not about the average stress in a material but about the energy balance at the tip of a crack. A crack will grow when the strain energy released by its extension exceeds the energy required to create new surfaces. Below that threshold, the crack sits dormant. Above it, it propagates — and in brittle materials under the right loading conditions, almost instantaneously.

Griffith’s insight was that strength is not a property of the material alone. It is a property of the material and its flaws. A sheet of glass is theoretically strong enough to support enormous loads. But a tiny scratch concentrates stress at the crack tip to levels thousands of times higher than the nominal load, and the glass shatters. The scratch does not weaken the material; it reorganises how force flows through it, creating a local intensity that the average measurement misses entirely.

The discipline made this concrete in 1954, when two de Havilland Comet airliners — the world’s first commercial jets — disintegrated in flight. Investigators found that the square corners of the passenger windows concentrated stress far beyond what the overall fuselage load would suggest. Pressurisation cycles drove microscopic cracks at these corners — damage too small to detect on any individual flight, but accumulating irreversibly until a crack reached critical length and the structure failed catastrophically. The fuselage had been strong enough for any single pressurisation. It was not strong enough for thousands.

Two concepts from the discipline are particularly useful for our purposes. The first is the stress intensity factor — a measure of how much stress is amplified at the tip of a flaw given its geometry and the applied load. Two structures carrying identical average loads can have vastly different stress intensities at their critical points. The structure with the sharper flaw is closer to failure, even though nothing about the average load distinguishes them.

The second is fatigue: the process by which repeated loading below the failure threshold drives incremental crack growth. The Paris law — an empirical relationship describing crack growth in its stable mid-range regime — reveals that the process is self-reinforcing: as the crack grows, the stress intensity increases, which accelerates the growth rate, which lengthens the crack further. Long periods of apparently stable operation are followed by rapid acceleration and sudden failure.

Where Stress Concentrates

What does this framework reveal when applied back to portfolios?

Start with stress concentration. The standard tools of portfolio risk — volatility, correlation, value-at-risk — are aggregate measures. They describe the overall stress field. They do not identify the features that concentrate stress locally, the points where the experienced load far exceeds the nominal one.

In a portfolio, stress concentrations take several forms. Illiquid positions concentrate stress because they cannot shed load when the overall portfolio is under pressure. A portfolio with 5 per cent in an illiquid asset does not have 5 per cent exposure in a drawdown; it has whatever-is-left-after-selling-the-liquid-assets exposure, which can be far higher. The illiquidity is the sharp corner in the window — a geometric feature that reorganises how stress flows through the structure.

Leverage is another concentrator. A leveraged fund does not simply have more exposure; it has a feedback mechanism where losses trigger margin calls, which force sales, which create further losses. The leverage concentrates stress at the point of forced selling, creating local intensities that exceed what the nominal exposure would suggest. In the 1994 bond rout, this was the mechanism that turned a sequence of modest rate rises into fund-ending events: each move consumed capital, which tightened leverage constraints, which forced selling, which locked in losses that further consumed capital.

Funding mismatches concentrate stress similarly. A portfolio of long-dated assets financed with short-term borrowing has a stress concentration at the rollover point. Under normal conditions, the mismatch is invisible — the funding renews routinely. Under stress, it becomes the point of failure. This was the structure of Northern Rock, of the conduits and SIVs in 2007-08, and of many carry trades before and since. The ergodicity problem we explored in The Path The Maths Misses applies here directly: the ensemble average across many possible funding states looks manageable, but the time-average path through a funding crisis is not.

Invisible Damage

Now consider fatigue. Portfolios do not experience a single load and either hold or fail. They experience repeated loading cycles — drawdowns, recoveries, periods of underperformance, liquidity squeezes that come and go. Each cycle can leave invisible damage.

What does damage look like in a portfolio? It is the erosion of the buffers that allow the portfolio to absorb future stress. Capital consumed by losses. Liquidity reserves drawn down. Risk budgets exhausted. Organisational patience depleted. Each drawdown-and-recovery cycle that looks like the system returning to normal may in fact be the system returning to a slightly weaker version of normal — the crack a little longer, the stress intensity a little higher, the margin of safety a little thinner.

This connects to the stability framework we developed in What Makes Stable Things Break. There, we described how a system can appear stable while the basin of attraction around it is shrinking — the ball still sitting at the bottom of the bowl while the bowl gets shallower. Fatigue is one mechanism by which the basin shrinks. Each loading cycle erodes the structural reserves that define the basin’s depth, even as the system returns to what looks like the same resting state.

The Paris law dynamic has a portfolio analogue. As buffers erode, the portfolio becomes more sensitive to the next stress event — remaining positions more concentrated, liquidity cushion thinner, organisational willingness to hold through pain diminished. The damage from one cycle feeds into the severity of the next.

This is why portfolios can survive ten drawdowns and fail on the eleventh, even when the eleventh is no larger than the ones before it. The question is not how large the current stress is but how much damage has already accumulated. It was not the final pressurisation cycle that destroyed the Comet. It was every cycle before it.

The stress concentrations — illiquidity, leverage, funding mismatches, counterparty dependencies — are often in the least visible parts of the structure. In The Network You Cannot See, we explored how hidden connections create links between apparently independent positions. Those hidden links are precisely where fatigue damage accumulates undetected.

The standard risk report is a static strength test. It asks: can the portfolio survive this scenario? But it does not ask: how has repeated stress changed the portfolio’s capacity to survive the next scenario? It does not track the sub-critical damage.

The Fragility Question

What might a fracture-mechanics-informed approach look like? Not prediction — materials science does not predict when a specific component will fail. But it can change what you monitor. Map the stress concentrations — illiquid positions, leveraged positions, funding or counterparty concentration — and give them scrutiny disproportionate to their weight. Track cumulative damage, not just current state: is capital being rebuilt after drawdowns, or is each recovery incomplete? And distinguish between damage-tolerant structures — diversified, liquid, modestly leveraged — and damage-intolerant ones — concentrated, leveraged, dependent on a single thesis or a single source of funding. The Comet disasters accelerated a shift in aircraft design toward damage-tolerant philosophy: assume cracks will form, design so that they grow slowly enough to be detected, and build in redundant load paths so that no single crack can bring down the structure. The difference matters enormously under repeated loading, even if both structures look adequate under a single static test.

The framework extends beyond finance. In organisations, repeated restructurings can fatigue institutional capacity even when each individual restructuring appears successful. The formal structure recovers; the informal networks accumulate damage that is not repaired by the next reorganisation. In infrastructure, repeated minor flooding events can degrade foundations, embankments, and drainage systems in ways that each post-event inspection clears as sound. The recovery looks complete from the outside. The damage is internal and cumulative.

The discipline is in asking: not whether something can survive this shock, but what has every previous shock already left behind?

References & Further Reading

1. Fracture Mechanics: Fundamentals and Applications: T.L. Anderson

2. Structures: Or Why Things Don’t Fall Down: J.E. Gordon

3. The Stress Analysis of Cracks Handbook: Hiroshi Tada, Paul Paris & George Irwin

4. Against the Gods: The Remarkable Story of Risk: Peter Bernstein

5. Big Bets Gone Bad: Derivatives and Bankruptcy in Orange County: Philippe Jorion

6. The (Mis)Behaviour of Markets: Benoit Mandelbrot & Richard Hudson

Beginning in October 1979, Paul Volcker’s Federal Reserve adopted a tightening regime that would push the federal funds rate above twenty per cent by early 1981. The objective was to break an inflationary spiral that had persisted for more than a decade. The cost was enormous: a deep recession, unemployment approaching eleven per cent, a wave of corporate and agricultural bankruptcies. The intervention was deliberate and blunt. Inflation was not declining because the economy had found a new equilibrium. It was declining because the Federal Reserve was suppressing demand with sufficient force to override the prevailing inflationary dynamics.

By any reasonable measure, the stability was imposed. The Fed was holding the economy in a state it would not have reached on its own, at a cost that was visible, escalating, and politically unsustainable. If the intervention had been withdrawn in 1980 or 1981, inflation would likely have re-accelerated — credibility had not yet been secured and the underlying dynamics had not changed.

But something happened over the following decade. Inflation expectations anchored. Wage-setting behaviour adapted to a low-inflation environment. Firms began pricing on the assumption that inflation would stay low, and their pricing decisions helped make it so. Central bank credibility — earned through the severity of the initial intervention — became a structural feature of the economy rather than an ongoing act of will. By the early 1990s, the Fed did not need to suppress the economy to keep inflation low. The system’s own dynamics were doing it.

The stability had been imposed. And then it took root.

For roughly three decades, this self-sustaining low-inflation equilibrium helped underpin the bond rally and, over time, the negative stock-bond correlation on which much multi-asset portfolio construction relied. It required far less effort to maintain than it had to establish — anchored expectations, credible institutions, and self-reinforcing behaviour did most of the work. Perturbations were absorbed. Oil shocks came and went. Recessions came and went. Each time, the system returned to low inflation without requiring Volcker-scale intervention.

What kind of problem is this?

When the question is not whether an equilibrium holds today but whether genuine stability has formed — when you need to distinguish between a system that is self-sustaining and one that is still being held in place by force that could be withdrawn — that is a specific shape of problem.

And the mathematical discipline that has formalised the conditions under which an equilibrium is self-restoring after disturbance is Lyapunov stability theory.

So let’s borrow.

This is the multi-model move: recognise the shape of a problem, find the discipline that has thought rigorously about that shape, and import its frameworks deliberately rather than reinventing them from scratch.

Lyapunov theory was built for mechanical and electrical systems — pendulums, circuits, control loops. Markets do not satisfy its assumptions cleanly: participants adapt, dynamics are reflexive, intervention is endogenous. But the structural logic — the habit of asking whether disturbances are being endogenously damped under the current regime — is close enough to our problem to borrow. The vocabulary for energy dissipation, equilibrium conditions, and the distinction between stability that is earned and stability that is purchased transfers.

The Function That Tells the Truth

Aleksandr Lyapunov, a Russian mathematician working in the 1890s, asked a deceptively simple question: how can you determine whether a system will return to equilibrium after a disturbance, without having to solve the full equations of motion? His answer was elegant. Find a scalar function — think of it as a measure of the system’s “energy” — that is positive away from equilibrium and that strictly decreases along the system’s trajectories. If such a function exists, the system is not merely stable but asymptotically stable: perturbations do not just stay bounded — they actively decay. The equilibrium attracts.

The power of the approach is that you do not need to trace every possible trajectory. You need to find the right function — the right measure of how much the system has been disturbed — and show that it is shrinking over time. If the function is shrinking, the disturbance is decaying. If it begins to grow, the system is moving away from equilibrium even if the state variables themselves appear calm. If it flatlines — neither growing nor shrinking — the situation is ambiguous: the system may be on the edge of stability or may need additional structure to settle.

The analogy to physical energy is deliberate. A ball in a bowl is at a stable equilibrium because its total mechanical energy — kinetic plus potential, the natural Lyapunov function for a mechanical system — decreases as friction dissipates its motion. The ball overshoots, oscillates, but each swing is smaller. The energy shrinks monotonically. The ball settles. If you remove the friction, the ball oscillates indefinitely — total energy is conserved, never decreasing, and the ball never settles, even though it keeps returning near the bottom. If you invert the bowl, the ball rolls away: the energy framework tells you immediately that the equilibrium is unstable.

From Imposed to Formed

The Volcker disinflation illustrates the most important transition the Lyapunov lens can diagnose: stability that begins as imposed and becomes genuine.

In 1979–82, the system’s internal disequilibrium was at its maximum — inflation expectations were deeply unanchored. Every perturbation — an oil price shock, a fiscal expansion, a wage negotiation — had to be met with further tightening. The displacement from equilibrium was not decreasing through the system’s own dynamics. It was being forced down by external intervention. This is the signature of imposed stability: the system requires continuous, costly effort to remain in its current state.

Over the following decade, the picture reversed. Inflation expectations had become self-reinforcing. The system’s internal disequilibrium — the degree to which expectations were unanchored, a useful proxy for displacement from equilibrium — was dissipating naturally. Consequently, the scale of external intervention required to maintain stability genuinely decreased after each shock. The 1990–91 recession, the 1997 Asian crisis, the 2001 dotcom bust: each produced economic stress, and each time inflation stayed anchored without extraordinary monetary force. The system was dissipating perturbations through its own dynamics. The stability had formed.

The practical question for an investor in the 2010s was not “is inflation low?” — it visibly was. The question was: has the low-inflation equilibrium genuinely formed, or is it still being imposed? The Lyapunov lens gave a clear answer: formed. Expectations were anchored. The self-reinforcing mechanism was operational. This mattered because it meant the equilibrium could be relied upon — not forever, but for as long as the structural conditions that produced it persisted.

The Floor That Never Took Root

In September 2011, the Swiss National Bank announced a floor on the EUR/CHF exchange rate at 1.20. The SNB declared it would enforce the floor “with the utmost determination” and was prepared to buy foreign currency “in unlimited quantities.” Capital flowed in on the assumption that the floor would hold.

For three years it did. But the Lyapunov diagnostic told a different story from the Volcker case. The cost of maintaining the floor was not decreasing. It was growing. The SNB accumulated foreign reserves from roughly 260 billion francs at end-2011 to over 510 billion by end-2014, and by early 2015 the required pace of intervention had become, in the SNB’s own later description, rapidly increasing. The system was not learning to sustain the equilibrium on its own. Market participants were not anchoring on EUR/CHF 1.20 in the way that firms had anchored on low inflation — they were leaning on the SNB’s commitment, and the leaning was getting heavier.

The stability was imposed. And it never took root.

On 15 January 2015, the SNB abandoned the floor without warning. EUR/CHF repriced by fifteen to thirty per cent intraday, depending on the venue — the biggest one-day fall of the euro against the franc in the pair’s history. Several foreign exchange brokerages became insolvent or required rescue. Funds that had treated the floor as an equilibrium suffered severe losses.

The contrast with the Volcker case is the diagnostic. Both began as imposed stability. In one case, the system’s own dynamics gradually took over — expectations anchored, behaviour adapted, the intervention could be withdrawn without the equilibrium collapsing. In the other, the system never developed self-sustaining dynamics. The intervention was the equilibrium. Remove it, and there was nothing underneath.

The structural logic suggests a useful extension: what we might call a stability budget. An imposed equilibrium consumes a budget continuously — the SNB’s budget was its institutional tolerance for balance sheet expansion, and that tolerance was being consumed at an accelerating rate. A genuinely formed equilibrium is not costless, but it is low-maintenance — the low-inflation regime after 1995 still depended on the implicit threat of future tightening, but it did not require Volcker-scale force to persist.

The transition from imposed to formed is, in these terms, the moment when the stability budget shifts from escalating to modest and episodic. The system moves from requiring large external force to dissipating perturbation energy largely on its own. Identifying that transition — or its absence — is the practical contribution of the lens.

The Formation Question

Lyapunov theory says nothing about timing. It cannot tell you when the SNB’s tolerance would run out, or when Volcker’s credibility would take hold. What it gives you is a structural test, and the test is specific.

Find the quantity that measures displacement from equilibrium. Not price. Not volatility. The internal state: how unanchored are expectations, how large is the intervention required to absorb a perturbation, how far is the system from a state it could sustain on its own. Then ask: is that quantity shrinking after each disturbance, or growing?

If it is shrinking through the system’s own dynamics, without increasing external force, stability is forming. You can rely on it, provisionally, for as long as the conditions that produce the self-correction persist. If it is shrinking only because something is paying to push it down, and the payments are getting larger, stability is imposed. You are relying not on the system but on the willingness of the imposer. And willingness is a decision, not a process. It can be withdrawn in an afternoon.

The test applies wherever stability is claimed. In professional standards, a norm that practitioners internalise and self-enforce is formed; a standard that holds only because of continuous external audit is imposed, and the distinction becomes visible the moment the auditor looks away. In organisations, a culture that self-reinforces through hiring and norms is formed; a culture that depends on a single leader’s force of personality is imposed, and the departure of that leader is the withdrawal of the floor.

The discipline is in asking: is the system’s displacement from equilibrium shrinking on its own, or is something paying to force it down?

References & Further Reading

1. The General Problem of the Stability of Motion: Aleksandr Lyapunov

2. Nonlinear Systems: Hassan Khalil

3. Stochastic Stability of Differential Equations: Rafail Khasminskii

4. Volcker: The Triumph of Persistence: William Silber

5. Thinking in Systems: A Primer: Donella Meadows

6. A History of Interest Rates: Sidney Homer & Richard Sylla

On October 15, 2014, the yield on the US 10-year Treasury note fell 16 basis points in six minutes, then nearly retraced over the next six. The day’s intraday range was 37 basis points — extraordinary for what is normally one of the most stable instruments in the world. No policy had changed. No institution had failed. After a modestly weak retail sales release, one-sided order flow met unusually thin depth and produced a sharp round-trip in the world’s deepest government bond market.

The Joint Staff Report published afterward found no single cause. One-sided order flow had overwhelmed the displayed depth. Principal trading firms sharply increased and then reversed their activity; bank-dealers intermittently withdrew. For a few minutes, the market’s capacity to absorb directional flow fell below the demand for immediate execution.

This was not a crisis. No one defaulted, no bank wobbled. But it revealed something important about liquidity. The standard mental model is that liquidity is a pool — a reservoir of available capital, deep or shallow. The pool metaphor is not wrong. But it is incomplete. Depth had already thinned before the event, but the pool had not emptied — capital had not fled the system. What had failed was the market’s ability to absorb one-sided flow at the rate it was arriving.

What kind of problem is this?

When a system functions smoothly under normal load but degrades sharply when demand spikes — not because resources have been exhausted but because absorptive capacity has been overwhelmed — that is a specific shape of problem. It is about the rate at which capacity can be delivered, not how much exists.

And the discipline that has spent a century studying how congestion emerges when variable demand meets finite capacity is queuing theory.

So let’s borrow.

This is the multi-model move: recognise the shape of a problem, find the discipline that has thought rigorously about that shape, and import its frameworks deliberately rather than reinventing them from scratch.

Queuing theory was built for telephone exchanges, motorways, and computer networks — systems with well-defined servers and orderly arrivals. A limit-order book is messier: a network of price-time priority queues with strategic participants who adapt in real time. But the structural vocabulary — for congestion, utilisation, and non-linear degradation — transfers.

The Mathematics of Waiting

Queuing theory began with a practical problem. In 1909, Agner Krarup Erlang, a Danish mathematician at the Copenhagen Telephone Company, needed to figure out how many circuits were required to handle the city’s call traffic.

Erlang’s insight was that average demand alone could not answer this. What mattered was the relationship between the arrival rate and the service rate — and crucially, the variability of both. Even if average capacity exceeded average demand, random clustering of arrivals could overwhelm the system in bursts.

This is the most counterintuitive thing about queuing theory. If ten calls arrive evenly spaced across ten minutes, no queue forms. If the same ten arrive in a burst during the first two minutes, the system is overwhelmed — even though average demand is identical. It is the clustering, not the volume, that creates the queue.

The simplest queuing model — the canonical M/M/1 queue, meaning random arrivals, random service times, one server — produces a result that is elementary and profound. Define the utilisation ratio, ρ, as the arrival rate divided by the service rate. When ρ is low, the system runs smoothly. As ρ approaches 1, average time in the system does not increase linearly. It bends upward sharply, approaching infinity.

At 50% utilisation, the system is comfortable. At 80%, time in the system has more than doubled. At 90%, it doubles again. At 95%, again. The relationship is convex — the last few percentage points of capacity matter enormously more than the first few. The result is robust across a wide family of single-server queuing models, though the exact curve depends on architecture.

The core intuition — that congestion grows convexly as utilisation rises — applies wherever variable demand competes for finite capacity: motorways, emergency departments, computer networks, and financial markets. The exact curve depends on architecture, but the non-linearity is general.

Where the Queue Forms

Now consider a financial market. Liquidity-demanding order flow, such as market orders and aggressive limit orders, arrives at some rate. The market’s ability to absorb it near prevailing prices depends on the displayed depth of the book and how quickly providers replenish it after trades occur. In a telephone exchange, the cost of congestion is time — you wait on hold. In a limit-order book, impatient flow converts waiting time into price impact.

In normal conditions, the service rate comfortably exceeds the arrival rate, and the system feels liquid. But arrival rates are not constant. They cluster. A macroeconomic release, a large rebalance, a sudden shift in sentiment — any of these can spike the arrival rate.

The problem comes when the spike in arrival rate coincides with a drop in service rate. And this is precisely what tends to happen, because the two are coupled. When order flow surges, market makers face adverse selection risk — the incoming flow may be informed. The rational response is to widen spreads or pull quotes, reducing absorptive capacity at the exact moment demand for it is highest.

This is the queuing insight applied to markets: they do not just experience higher arrival rates in stress. They experience higher arrival rates and lower service rates simultaneously.

Kyle’s 1985 model formalised why order flow moves price: market makers infer information from aggregate flow, so each unit of imbalance carries a price impact cost. That is not a queuing service rate, but it explains why the price of immediacy rises when flow clusters. Cont, Stoikov, and Talreja then modelled a stylised limit-order book as a queuing system, making the connection between microstructure and queuing theory direct.

Bouchaud, Farmer, and Lillo documented persistent dynamics in order flow — long memory in order signs, clustering in trade sizes, fluctuating depth — that make a queuing interpretation natural. Microstructure already has a name for this: resiliency, the speed at which depth replenishes and prices stabilise after a shock. The queuing framework makes that concept precise.

The Steep Part of the Curve

The queuing framework explains something that the pool metaphor alone cannot: why liquidity does not drain gradually. If liquidity were only a pool, you would expect it to deplete smoothly. Instead, what markets repeatedly demonstrate is an abrupt, non-linear shift: liquid conditions one moment, gridlock the next, with very little in between.

The October 2014 Treasury event showed this. So did the equity market on August 24, 2015, when impaired price discovery and reduced displayed depth produced severe dislocations in some exchange-traded funds — nearly a fifth fell 20% or more, far exceeding contemporaneous moves in their underlying baskets.

In each case, the system was not out of liquidity in the pool sense. What was missing was depth and replenishment. The queue had entered the steep part of its curve.

This connects to the ergodicity problem from The Path The Maths Misses. The ensemble average — “on average, markets are liquid” — may be true and yet irrelevant to the investor who needs to execute during the minutes when they are not. Edge, as we defined it in The Question Shapes The Answer, is conditional information — but information you cannot act on is not edge at all. The queue is the bottleneck between knowing and doing.

It also connects to The Network You Cannot See. In a network of queues, congestion propagates. When one venue hits its capacity limit, order flow diverts to related markets, potentially overwhelming their absorptive capacity in turn. The network determines where congestion starts; the queuing dynamics determine how fast it escalates.

Slow In, Fast Out

There is a deeper asymmetry. In most markets, the agents providing absorptive capacity — market makers, dealers — have few binding obligations to do so. Some venues have designated market makers, but those duties are typically weak and waivable under stress. In practice, liquidity is provided when profitable and withdrawn when not. The service rate is endogenous.

In Erlang’s telephone exchange, the circuits did not unplug themselves when call volume spiked. In financial markets, the service rate is a strategic variable. When conditions deteriorate, the rational response for a liquidity provider is to reduce capacity — exactly the response that pushes the utilisation ratio toward the critical zone.

This creates a structural asymmetry. Liquidity accumulates gradually as market makers compete for flow, narrowing spreads and deepening books over weeks and months. But it can withdraw in seconds as providers simultaneously pull back. The buildup is slow; the collapse is fast.

The same asymmetry appears wherever service capacity is voluntarily provided. In The Room That Runs Out, we saw how carrying capacity constrains populations. The queuing lens adds a mechanism: the room shrinks precisely when the crowd grows.

The Capacity Question

Queuing theory does not predict when a liquidity event will occur. What it does is add a question. The pool question — “how much liquidity is there?” — is necessary but not sufficient. The queuing question asks: what is the relationship between the demand for liquidity and the capacity to supply it, and how does that relationship change under stress?

The framework asks specific things. How close is the demand for liquidity provision to the supply of it? What would cause the arrival rate to spike — concentrated positioning, correlated rebalancing flows? What would cause the service rate to drop — adverse selection, balance sheet constraints, concentration of provision in a few firms? And critically: are the two likely to move in the same direction under stress?

These questions apply beyond financial markets. In healthcare, emergency departments operate on a version of the same curve — operational research suggests that hospitals above roughly 85% bed occupancy begin losing surge capacity, and performance degrades sharply as occupancy climbs toward the mid-nineties. In organisations, decision-making bottlenecks follow the same logic: an approvals process that works at moderate load can seize when demand clusters.

The discipline is in asking: not how much capacity exists but can it be delivered at the rate you need it and when you need it most?

References & Further Reading

1. The Theory of Probabilities and Telephone Conversations: Agner Krarup Erlang

2. Continuous Auctions and Insider Trading: Albert S. Kyle

3. A Stochastic Model for Order Book Dynamics: Ramy Cont, Sasha Stoikov & Rishi Talreja

4. How Markets Slowly Digest Changes in Supply and Demand: Jean-Philippe Bouchaud, J. Doyne Farmer & Fabrizio Lillo

5. The Joint Staff Report: The U.S. Treasury Market on October 15, 2014: US Treasury, Federal Reserve, SEC, CFTC

6. Market Liquidity: Theory, Evidence, and Policy: Thierry Foucault, Marco Pagano & Ailsa Röell

In the mid-1980s, a small number of trading firms discovered that the price of S&P 500 futures frequently diverged from the value of the underlying basket of stocks. The gap could be wide — sometimes several index points — and the trade was mechanical: buy the cheap side, sell the expensive side, wait for convergence. The maths was straightforward, the convergence was guaranteed at expiry, and the early practitioners earned extraordinary returns for what was, in principle, a riskless arbitrage.

Within a few years, the niche had transformed. Dozens of firms built the infrastructure to execute the same trade — direct exchange connections, faster execution systems, dedicated capital. By the early 1990s, the spreads had compressed from several index points to fractions of a basis point. The trade still existed. The convergence was still guaranteed. But the returns had collapsed to levels that could barely cover the cost of the technology required to capture them.

The index arbitrageurs had not been defeated by a change in market structure or a shift in the rules. The futures still diverged from the basket; the basket still converged at settlement. What had changed was the population. Too many firms were chasing the same spread, and the spread could not sustain them all.

What kind of problem is this?

The core opportunity has not disappeared. The spread still exists. The problem is that too many participants are competing in the same territory, and the ecosystem cannot support them all. When returns compress not because the opportunity has moved but because the capital chasing it has multiplied, that is a population dynamics problem. And the discipline that has spent decades studying what happens when populations exceed their resource base is ecology.

So let’s borrow.

This is the multi-model move: recognise the shape of a problem, find the discipline that has thought rigorously about that shape, and import its frameworks deliberately rather than reinventing them from scratch.

Ecology studies organisms that do not read about themselves or change strategy in response to being studied. Markets are faster, noisier, and more reflexive. But the structural vocabulary — for competition, saturation, and differentiation — is worth importing.

The Hypervolume

A niche, in ecology, is not just a place. It is a set of conditions and resources that allow a species to persist. The ecologist G. Evelyn Hutchinson formalised this in 1957 as the “n-dimensional hypervolume” — the full set of environmental conditions within which a species can survive and reproduce. The fundamental niche is the range of conditions a species could exploit in the absence of competitors. The realised niche is the smaller range it actually occupies once competition and other constraints are accounted for.

Carrying capacity — typically denoted K — is the maximum population size that an environment can sustain given available resources. It is not a fixed ceiling; K shifts as conditions change. But the logistic growth model captures the core dynamic: growth is fastest when numbers are low relative to resources, and decelerates as the population approaches K. The resource does not disappear. It simply cannot support unlimited extraction.

Competitive exclusion — Gause’s principle — states that two species competing for exactly the same niche cannot coexist indefinitely. The resolution is typically extinction or niche partitioning: selective pressures drive the species to specialise on slightly different resources, reducing direct competition. Spatial and temporal variation can sustain coexistence even with significant overlap, but where stabilising differences are weak, exclusion is the limiting case.

What Alpha Depletes

Consider the niche first. An investment strategy occupies a niche defined by the conditions that make it profitable: a particular type of mispricing, a specific market structure, a set of behavioural regularities that generate exploitable patterns. Merger arbitrage occupies the niche of announced-deal spreads. Statistical arbitrage occupies the niche of short-term mean reversion in correlated securities. Distressed debt occupies the niche created by forced sellers and information asymmetry in bankruptcy.

Each niche has a carrying capacity — not a fixed number, but a moving threshold set by the amount of capital it can absorb before returns degrade. In ecology, the relevant population measure is biomass; in finance, the equivalent is deployed capital. The threshold depends on the depth and frequency of the opportunities, the transaction costs, and the degree to which capital inflows themselves alter the dynamics. But it is finite.

This is where the ecology lens adds something. “Alpha decay” is conventionally framed as a property of the strategy — as though strategies age and weaken like radioactive isotopes. But the ecological view suggests it is not primarily about the strategy. It is about the relationship between the strategy and the population exploiting it. A strategy does not decay in isolation. It decays when the niche fills up.

The information-theoretic framework from The Question Shapes the Answer reinforces this. Edge is relational — it exists between an observer and a market, not as an intrinsic property of the observer. The ecological lens adds a further dimension: edge is also population-dependent. A signal that is highly informative when you are the only one trading it becomes noise when a thousand funds do the same thing.

The logistic growth model maps directly. When a new strategy is discovered, early entrants earn high returns. Capital flows in. Each new entrant captures a share of the finite return pool. Returns compress. Some funds close. Others persist with diminished returns. The opportunity still exists — but the population has reached the level the resource can sustain.

Unlike mineral deposits, mispricings are a renewable resource — continuously regenerated by hedgers, index rebalancers, and liquidity-seeking flows. But carrying capacity is set by the rate of renewal, not the existence of the resource. When extraction by arbitrageurs exceeds the rate at which the market regenerates the mispricing, returns compress even though the opportunity never disappears.

Surviving Past Zero

Competitive exclusion applies with a twist. In biology, species competing for identical resources cannot coexist — one will eventually dominate. In finance, funds competing for identical alpha cannot all earn excess returns — but they can all survive at zero net alpha if subsidised by management fees. The niche can remain overcrowded long after carrying capacity has been exceeded, because the population is decoupled from the resource it nominally exploits. Fees, not returns, sustain it.

This has implications for how we evaluate strategies. The question is not simply “does this strategy have alpha?” but “what is the carrying capacity of this niche, and how much of it is already occupied?” A strategy can be logically sound, historically validated, and currently unprofitable — not because it is wrong, but because the niche is full.

Why Good Strategies Stop Working described how strategies decay through environmental mismatch — the ice melting beneath the polar bear. The ecology lens identifies a second mechanism: the niche filling up with polar bears. Both produce declining returns but demand different responses. Environmental mismatch calls for adaptation. Niche saturation calls for differentiation. Confusing the two leads to the wrong intervention.

Beaks and Seeds

Competitive exclusion sounds terminal, but ecology offers a resolution. When two species face exclusion from the same niche, they can survive by partitioning it — specialising to exploit a subset of the resource that the other does not. Darwin’s finches are the canonical example. Thirteen species coexist on the Galapagos Islands, all descended from a common ancestor, by evolving beaks suited to different food sources. One cracks hard seeds. Another probes for insects. A third feeds on cactus flowers. They occupy the same archipelago but different ecological niches. The resource is shared; the means of extraction are distinct.

Niche partitioning is the mechanism by which ecosystems sustain diversity. Without it, competitive exclusion would drive every crowded habitat toward a single dominant species. With it, multiple species coexist by carving the resource space into territories narrow enough that direct competition is reduced. The narrower the specialisation, the more species the ecosystem can support.

In investing, this explains why “the same strategy” can have very different outcomes for different practitioners. Two quantitative equity funds may both call themselves “momentum” investors, but one trades weekly signals in large-cap equities while the other trades monthly signals in emerging markets. They occupy different niches despite sharing a label. Successful long-lived investment firms tend to be those that have found niches with structural barriers to entry — capacity constraints that limit competition, informational advantages that are costly to replicate, or time horizons that most capital cannot tolerate. The ergodicity problem from The Path the Maths Misses creates one such barrier: strategies that require tolerance for deep drawdowns and long recovery periods are inaccessible to capital with short evaluation horizons, even if the long-run returns are attractive. The inability to survive the path is itself a carrying-capacity constraint on the competition.

In a world of rapid information dissemination and freely available data, easily replicable strategies reach carrying capacity faster than ever. Strategies that require patient capital, deep expertise, or tolerance for illiquidity face less competition precisely because the barriers to entry are higher.

The Ecosystem Question

When returns are compressing and the debate is “does this strategy still work?”, the answer may be “yes, but the niche is full.”

The framework asks specific questions. What niche does this strategy occupy? What is the carrying capacity — how much capital can it absorb before returns degrade? How occupied is the niche currently? And is there a path to partitioning — a related but distinct opportunity that the crowd has not reached?

The ecology lens does not tell you when to enter or exit a strategy. It does not predict which niches will fill or when new ones will open. What it does is provide a structural vocabulary for familiar phenomena. “Alpha decay” becomes carrying capacity. “Crowding” becomes competitive exclusion in progress. “Finding a new edge” becomes niche partitioning.

The pattern extends well beyond investing. In organisational strategy, it applies to business models that once occupied white space and now face dozens of competitors extracting the same value. In scientific research, new subfields follow the same arc: early papers have high impact, grant funding flows in, more researchers enter the field, and the marginal contribution per paper declines even as the total output grows. Wherever success attracts imitation, and imitation compresses the reward for the original insight, the carrying capacity question applies. The room runs out in every domain.

Understanding the ecology of your position does not make you immune to competition. You are not the ecologist observing the ecosystem from outside. You are one of the organisms in it, subject to the same dynamics of crowding and exclusion that the framework describes. The finches that understood carrying capacity would still have to eat.

The discipline is in asking: where do I fit in this ecosystem?

References & Further Reading

1. Concluding Remarks (The Niche Concept): G. Evelyn Hutchinson

2. The Struggle for Existence: G. F. Gause

3. Adaptive Markets: Financial Evolution at the Speed of Thought: Andrew Lo

4. Frontiers of Finance: Evolution and Efficient Markets: Farmer & Lo

5. Mutual Fund Flows and Performance in Rational Markets: Berk & Green

6. The Rate of Return on Everything, 1870-2015: Jorda, Knoll, Kuvshinov, Schularick & Taylor

An investment committee reviews one hundred ideas a year and approves twenty. Of those twenty, twelve deliver returns. The committee reports a 60% hit rate. Not bad.

But is it good? You cannot answer that question with the information given. Suppose the pipeline contained eighty genuinely attractive ideas and twenty poor ones. The committee approved twelve that delivered and eight that did not — but also rejected sixty-eight that would have delivered. Its 60% hit rate conceals a significant miss rate. Alternatively, suppose only fifteen of the hundred ideas were genuinely attractive, and the committee found twelve of them. Now the same 60% hit rate reflects considerable skill. In practice, rejected ideas vanish — you rarely learn whether they would have succeeded. But shadow portfolios and post-hoc tracking of passed opportunities suggest the problem is real, even when the exact numbers are unknowable.

The hit rate isn’t a measure of skill. It’s a tangled composite of two completely different things: how well the committee can distinguish good ideas from bad ones, and how willing it is to say yes. These aren’t the same thing.

What kind of problem is this?

When you are trying to detect something real against a background of noise, and your detection system is imperfect, that’s a classification problem — not a forecasting problem or an optimisation problem. Is this thing I’m looking at real, or is it noise? And the discipline that has formalised the tradeoffs, the errors, and the hidden role of base rates is signal detection theory.

So let’s borrow.

This is the multi-model move: recognise the shape of a problem, find the discipline that has thought rigorously about that shape, and import its frameworks deliberately rather than reinventing them from scratch.

Signal detection theory (SDT) emerged from a collaboration between psychophysics and electrical engineering in the 1950s, codified in Green and Swets’ foundational 1966 text. The original context was literal: radar operators in the Second World War, staring at screens, trying to decide whether each blip was an enemy aircraft or atmospheric noise. Real signals and random noise produced overlapping patterns, and no amount of training could eliminate the ambiguity.

Sensitivity and Criterion

SDT formalises this by imagining two overlapping probability distributions — one for noise alone, one for signal-plus-noise. The overlap is the entire problem. The framework’s central insight is that detection performance decomposes into two independent dimensions.

The first is sensitivity — how far apart the two distributions are. In the technical language, this is d-prime (d’), the standardised distance between the means of the noise and signal distributions. High sensitivity means the world looks genuinely different when a signal is present versus when it isn’t. Low sensitivity means the two states are nearly indistinguishable. Sensitivity is a property of the detection system. You improve it by getting better information, better models, better analytical tools.

The second is the criterion — the threshold at which the observer decides to say “signal.” This is a choice, not a measurement. Slide the criterion leftward (become more liberal) and you catch more genuine signals but also flag more noise. Slide it rightward (become more conservative) and you reduce false alarms but miss more genuine signals. At any given level of sensitivity, this tradeoff is inescapable. The only way to get more hits without more false alarms is to improve sensitivity itself.

Every detection decision produces one of four outcomes: a hit (signal present, correctly identified), a miss (signal present, overlooked), a false alarm (noise mistaken for signal), or a correct rejection (noise correctly dismissed). Note that SDT’s “hit rate” — the percentage of genuine signals you caught — is not the same as the usage from our opening, where “hit rate” meant the percentage of approvals that worked out. These are different numbers, and confusing them is itself a diagnostic error. Plot the true hit rate against the false alarm rate as you sweep the criterion, and you trace the ROC curve — the Receiver Operating Characteristic. A system with high sensitivity bows toward the upper left; one with zero sensitivity produces a diagonal line. The area under the curve is a threshold-invariant measure of discrimination ability, untangled from any particular criterion setting.

This decomposition is SDT’s deepest contribution. In most domains, performance is evaluated as a single number — accuracy, hit rate, the percentage of investments that worked. But a single number conflates two things that should be assessed separately. A conservative investor who approves few ideas and gets a high percentage right might be cautious, not skilled. A liberal investor who approves many and gets a lower percentage right might have genuine sensitivity combined with a willingness to act. You can’t distinguish these cases without pulling the two dimensions apart.

The Base Rate Trap

Suppose genuine investment opportunities — ones that would generate meaningful risk-adjusted returns — constitute 5% of the ideas crossing an investment committee’s desk. The committee has good detection: when a genuinely good idea appears, they recognise it 80% of the time, and they correctly reject 90% of bad ideas. Strong numbers. Skilled committee.

Now do the arithmetic. Out of 1,000 ideas, 50 are genuinely good. The committee identifies 40 of them. Of the 950 bad ideas, it incorrectly approves 95. Total approvals: 135, of which 40 are good. The positive predictive value — the probability that an approved idea is actually good — is roughly 30%.

A 30% success rate from a highly skilled committee. Not because the committee is bad, but because the base rate is low. When genuine signals are rare, even excellent detectors generate mostly false positives. This isn’t an argument against trying. It’s an argument for understanding what the numbers actually mean.

The Payoff Matrix

In The Question Shapes The Answer, we explored Adami’s framework for information: edge requires specifying a target, an observer, and a baseline. SDT adds a complementary layer. Even when you have genuine conditional information — even when your sensitivity is real — the decision to act on that information involves a separate, strategic choice about where to set your threshold. And that choice should depend on the payoff structure and the base rate of genuine opportunity, not on the signal itself.

Consider the asymmetry. For a long-term investor with a diversified portfolio, the cost structure might look like this: missing a genuinely great investment is expensive because opportunity cost compounds over decades, while an underperforming investment drags returns but doesn’t threaten long-term viability. This cost structure argues for a liberal criterion — approve more, accept that many will underperform, because the cost of missing strong investments exceeds the cost of including weaker ones.

For a concentrated, leveraged fund, the cost structure inverts. A false alarm — a position that looked like a signal but was noise — can be catastrophic. The ergodicity problem from The Path The Maths Misses applies: a single bad position can destroy the ability to compound. This argues for a conservative criterion — demand compelling evidence before acting, accept that you will miss many genuine opportunities, because survival dominates.

Both are rational. Neither is “better.” They reflect different payoff matrices applied to the same underlying detection problem. SDT suggests that framing high-conviction and diversified approaches as inherently superior or inferior is a category error. The question isn’t how much conviction to have. It’s what the costs of your errors are, and whether your criterion is calibrated to those costs.

The Invisible Threshold

This connects to something the evolutionary lens from Why Good Strategies Stop Working revealed: investment strategies carry assumptions about their environment. SDT sharpens this. Every organisation carries criterion — a collective threshold for what counts as a sufficient signal to act. It is often implicit, embedded in how many approvals the committee gives, how much evidence is required, what the burden of proof feels like.

Some organisations are miss-averse and set liberal criteria. Others are false-alarm-averse and set conservative criteria. These aren’t personality quirks. They are strategic positions in the sensitivity-criterion space. But they are often invisible — embedded in norms rather than articulated as choices.

The only way to improve both simultaneously is to improve sensitivity — to actually get better at distinguishing signal from noise. If you decompose performance into sensitivity and criterion, you can ask the right question: is our problem that we are poorly calibrated (criterion in the wrong place), or that we genuinely cannot tell signal from noise (low sensitivity)? These require completely different interventions. Recalibrating a criterion is a decision. Improving sensitivity is a capability-building exercise — better data, better models, better analytical infrastructure, the kind of investment in measurement apparatus that The Question Shapes The Answer described.

Post-mortems often confuse the two. When an investment fails, the typical response is to raise the bar — demand more evidence, add another approval layer. This is a criterion shift. If the problem was low sensitivity — if the system genuinely could not distinguish signal from noise — then raising the bar simply means you will miss more signals while still getting fooled by noise that exceeds your new threshold. You have made yourself more cautious without making yourself more skilled. The ROC curve has not moved; you have merely walked along it.

The Detection Question

The discipline SDT offers is in the decomposition. Before asking “should we approve more or fewer ideas,” ask: can we actually tell the difference? Before tightening standards after a loss, ask: was the failure a criterion problem or a sensitivity problem? Before celebrating a high hit rate, ask: what was the base rate, and how many signals did we miss?

In markets, this decomposition interacts with competition. When everyone is trying to detect the same signals, the base rate of genuinely private information drops — more observers means more information gets priced faster. The more efficient the collective detection system, the lower the base rate for any individual. Improving aggregate sensitivity reduces each participant’s base rate, which means even skilled detectors will find that most of their positive identifications are false alarms. Understanding signal detection theory doesn’t make you immune to its dynamics. You will still face ambiguous signals, still set your criterion imperfectly, still be surprised by how often confident calls turn out to be noise.

The decomposition applies wherever imperfect detectors face ambiguous signals. In medicine, a screening test with excellent sensitivity can still produce mostly false positives when the disease is rare — and the response of ordering more tests is a criterion shift, not a sensitivity improvement. In security, tightening protocols after an incident can flood the system with false alarms while doing nothing to improve the ability to detect genuine threats. Any domain where someone is trying to sort signal from noise — and where the cost of errors is asymmetric — is a domain where SDT’s decomposition reveals something that a single accuracy number conceals.

But the framework changes the questions you ask. Not “was I right?” but “can I actually distinguish signal from noise in this domain?” Not “how many decisions worked?” but “what was the base rate, and does my sensitivity exceed chance?”

The discipline is in asking: is this a problem of calibration or a problem of capability?

References & Further Reading

1. Signal Detection Theory and Psychophysics: David M. Green and John A. Swets

2. The Signal and the Noise: The Art and Science of Prediction: Nate Silver

3. Noise: A Flaw in Human Judgment: Daniel Kahneman, Olivier Sibony, and Cass R. Sunstein

4. Thinking, Fast and Slow: Daniel Kahneman

5. The Base Rate Book: Integrating the Past to Better Anticipate the Future: Michael J. Mauboussin

6. What is Information?: Christoph Adami

In April 1925, Winston Churchill, then Chancellor of the Exchequer, returned Britain to the gold standard at the pre-war parity of £1 to $4.86. The model behind the decision was explicit: restoring the pre-war rate would signal stability, anchor expectations, and re-establish London as the centre of global finance.

The problem was that Britain was not what it had been. The economy was weaker, export industries were less competitive, and the real value of sterling at the old parity was roughly 10-12% too high. John Maynard Keynes said so almost immediately, publishing The Economic Consequences of Mr Churchill that summer. The pressures he predicted — deflation, wage compression, acute stress in export industries — all materialised.

What followed was six years of the British government acting on the economy to make it fit the model. The parity had to hold, so wages had to fall. Exports were uncompetitive, so costs had to be squeezed. The coal industry could not compete at the prevailing exchange rate, so miners faced wage cuts and lockouts — precipitating the General Strike of 1926. Unemployment was chronic. Deflation ground on. At every stage, the response to mounting evidence that the parity was wrong was not to revise the model but to intensify the intervention: tight money, credit restriction, wage pressure, internal devaluation. The economy was being reshaped to fit the exchange rate, rather than the exchange rate being adjusted to fit the economy.

In September 1931, Britain abandoned the gold standard. The adjustment that the model had suppressed for six years arrived all at once.

What kind of problem is this?

When the response to contradictory evidence is not to revise the model but to intervene harder to preserve it, when the cost of that preservation escalates until it breaks discontinuously, that is a specific shape of problem. Behavioural finance would call it bias. But there is a more precise framework, one that explains not just that this happens but why.

The discipline that has formalised this is the science of self-organising systems — and the framework is Karl Friston’s free energy principle, originally developed in theoretical neuroscience but applicable to any system that maintains itself over time against disorder: institutions, strategies, and markets as much as organisms and brains.

So let’s borrow.

This is the multi-model move: recognize the shape of a problem, find the discipline that has thought rigorously about that shape, and import its frameworks deliberately rather than reinventing them from scratch.

The standard account would call Churchill’s Treasury irrational — anchored to a sunk cost, trapped by confirmation bias, unable to update. These labels are real but they are descriptions, not explanations. They name the pattern without explaining why the architecture produces it so reliably. Friston’s framework says something more unsettling: the resistance to updating is not a malfunction. It is what self-organising systems are designed to do.

Two Ways to Be Less Surprised

In 2006, Friston proposed that any self-organising system that persists over time does so by minimising something he called variational free energy, a quantity from statistical physics that bounds surprise. There are two ways to reduce it. You can change the world to match your model: that is action. Or you can change your model to match the world: that is perception. Both narrow the gap between prediction and reality. The decision-making framework that follows — active inference — asks how a system chooses between possible courses of action, which Friston calls policies. The expected free energy of a policy can be read two ways.

The first reading decomposes it as:

Minimising G means maximising both. Pragmatic value captures how likely a policy is to produce outcomes the system prefers: acting on the world to make it conform to the model. Epistemic value captures how much a policy will reduce uncertainty about what is going on: seeking information that updates the model to better fit the world.

The second reading decomposes the same quantity as:

Risk is the divergence between predicted outcomes and preferred outcomes: how far the consequences of a policy fall from what the system wants. Ambiguity is uncertainty about outcomes given the state of the world: how noisy the mapping from reality to observations will be.

These are two readings of the same quantity.

There is a separate source of resistance to updating. Variational free energy, the quantity the system minimises when revising beliefs about the current state, decomposes into accuracy minus complexity. Complexity is the divergence between updated beliefs and prior beliefs. Moving your beliefs far from where they started incurs a cost, even if the new beliefs are more accurate. A system will patch an existing model before it bears the cost of replacing it.

The balance between action and perception is governed by what Friston calls precision — the inverse variance that weights how strongly prediction errors update beliefs. High precision on prior preferences means the system trusts its model and acts on the world. High precision on incoming evidence means prediction errors get amplified and the system updates readily.

A thermostat is all pragmatic value and no epistemic value. It does not revise its model of what the temperature should be; it acts on the room to match the set point. Churchill’s Treasury was a thermostat — forcing the economy to conform to a pre-war exchange rate rather than revising the rate to fit the economy.

The Precision Trap

This is where the framework bites. Both channels of resistance compound.

The gold standard commitment was not just an economic policy. It was national prestige made concrete. Abandoning the parity meant outcomes the Treasury found intolerable: political humiliation, loss of credibility, an admission that six years of policy had been wrong. And the complexity cost of revising the model was enormous — the beliefs had been held publicly, at the highest levels, for years. Both channels locked the system in place.

Each year the parity held made it harder to abandon. Not because the evidence improved — it worsened steadily — but because the cost of admitting error grew with every budget, every speech, every policy justified by the model’s persistence. The Treasury was not ignoring the evidence. It was processing it through a system whose precision settings made action cheaper than revision. Holding, averaging down, reinterpreting. These are the low-cost moves. Not from irrationality. From architecture.

This is also why crowded trades persist beyond the point where the evidence has turned. In What Makes Stable Things Break, we described how consensus views function as attractors — states the system gravitates toward, reinforced by positioning and shared belief. The free energy principle offers a mechanism: the consensus is a shared model with very precise prior preferences. Each participant is individually minimising free energy, and for each the lowest-cost path is to preserve the existing model. The crowd does not coordinate its stubbornness. The stubbornness emerges from each individual’s precision settings.

When the Update Finally Comes

But models do break. As the gap between model and reality widens, the cost of defending the model through action keeps rising. The expected free energy of the defend-the-model policy climbs. Eventually it exceeds the expected free energy of revising the model. When it does, the update arrives suddenly.

The transition is not gradual because the model was not gradually wrong. In 1925, and 1926, and every year after, the contradictory evidence was arriving — Keynes had published it, the unemployment data confirmed it, the General Strike dramatised it — but it was being met with intervention, not revision. The cost of the defend-the-parity policy was rising the entire time. When the dam broke in September 1931, all the accumulated error arrived at once. Sterling lost a quarter of its value. Export competitiveness improved almost immediately.

The pattern echoes across decades. On Black Wednesday in 1992, the British government raised interest rates twice and spent billions defending sterling’s peg to the deutschmark within the European Exchange Rate Mechanism. The model said sterling belonged in the band. The fundamentals, distorted by German reunification, said otherwise. The response was the same as in 1925: act on the world, not the model. When the defence failed that evening, the suppressed prediction error arrived all at once. Soros was not betting that the fundamentals were wrong — that was widely known. He was betting that the government’s precision on the peg was unsustainable, and that when it broke, the repricing would be violent.

The yen carry unwind of August 2024, which we examined in What Makes Stable Things Break, had the same character. The model — “the differential persists, the yen stays weak” — had been accumulating contradictory evidence. But precision was high: it had worked for years, and the cost of updating was real. When the update arrived, it was not a gentle revision but a cascade.

This is not irrationality punctuated by sudden rationality. It is the natural dynamics of a system that minimises free energy through action until action is no longer sufficient. In that earlier post, we called this a bifurcation. The free energy principle offers a mechanism for why systems hover near that threshold: the agents inside are actively working to prevent the crossing.

The Prediction Error Question

The framework applies wherever a model has accumulated institutional weight. Strategy reviews that always conclude the strategy is sound. Risk committees that hear the warning but reinterpret the data. Organisations that restructure to protect a thesis rather than test it. These are prediction-error-minimising systems doing what they are designed to do.

The practical discipline is not “be less biased” — that advice is empty. It is in monitoring both channels. How precise are your preferences over outcomes, and is that precision earned or accumulated through commitment? How large is the complexity cost of revision, and is it distorting your weighting of the evidence?

The asymmetry matters. A small, recent, lightly held position has low precision and will update easily. A large, long-held, publicly defended position has high precision and will resist until the evidence is overwhelming — at which point the update will be violent. The longer the dam holds, the larger the flood.

After Britain left the gold standard in 1931, the economy recovered faster than almost anyone expected. Exports grew, unemployment fell, industrial production rose. The adjustment the model had suppressed for six years took months once allowed to operate. The prediction error had been there all along. The system had simply been working very hard not to feel it.

The discipline is in asking: is the model holding because the evidence supports it, or because the cost of updating it is too high?

References & Further Reading

1. A Free Energy Principle for the Brain: Karl Friston, James Kilner & Lee Harrison

2. Action and Behavior: A Free-Energy Formulation: Karl Friston, Jean Daunizeau, James Kilner & Stefan Kiebel

3. Active Inference and Epistemic Value: Karl Friston, Francesco Rigoli, Dimitri Ognibene, Christoph Mathys, Thomas Fitzgerald & Giovanni Pezzulo

4. Active Inference: A Process Theory: Karl Friston, Thomas FitzGerald, Francesco Rigoli, Philipp Schwartenbeck & Giovanni Pezzulo

5. Surfing Uncertainty: Prediction, Action, and the Embodied Mind: Andy Clark

6. The Economic Consequences of Mr Churchill: John Maynard Keynes

7. The Alchemy of Finance: George Soros

In the winter of 1636, a speculative frenzy in tulip bulb futures swept through trading circles in Holland. In the best surviving price records, some varieties rose roughly tenfold in weeks. The bulbs themselves were in the ground — no one could see or enjoy them. What was being traded were contracts: promises to deliver bulbs in the spring. Contemporaries derided it as windhandel — wind trade — because contracts changed hands while the bulbs stayed in the ground. The object of speculation was not a flower. It was a claim on what someone else would pay for a claim on a flower.

Most participants were not horticulturists. They were merchants and tradesmen gathering in tavern trading rooms called collegia, watching prices rise and acting on the visible behaviour of others doing the same. The question driving the market was not “what is this bulb worth?” The question was “what will someone else pay for this contract tomorrow?”

On the first Tuesday of February 1637, at a routine auction in Haarlem, the auctioneer could not find willing buyers. Over the following days, the failure spread from town to town.

What kind of problem is this?

When participants must anticipate each other’s anticipations — when the “right” answer depends on what others believe about what others believe — that is a coordination problem dressed in financial clothing.

The standard valuation framework answers the question: “What is this asset worth, given its future cash flows?” That is meaningful when the asset has future cash flows. A tulip bulb in frozen ground does not. But the collegia were full anyway — because the participants were not pricing flowers. They were pricing each other’s intentions.

Tulips make the coordination structure visible because there is nothing else to see. But the same dynamic operates in any market — it is simply easier to overlook when the asset has real cash flows to distract you.

The question that matters is: “What do others believe, and what do others believe others believe — and what would change that?”

The discipline that has thought rigorously about this shape is game theory. Morris and Shin formalised how coordination failures emerge under higher-order uncertainty. But the sharpest intuitive articulation came earlier, from Keynes in the 1930s.

So let’s borrow.

This is the multi-model move: recognize the shape of a problem, find the discipline that has thought rigorously about that shape, and import its frameworks deliberately rather than reinventing them from scratch.

The standard framing asks what an asset is worth. The beauty contest asks what others believe others will pay — a different question with a different answer.

The Original Beauty Contest

In 1936, Keynes described a newspaper competition common in his era. Readers were shown photographs of faces and asked to pick the six most beautiful. The winner was whoever’s choices most closely matched the most popular choices overall.

The naive strategy is to pick the faces you find most beautiful. But this ignores the game. You are rewarded not for your aesthetic judgment but for anticipating the crowd’s.

So you move to second-order thinking: pick the faces you think others will find beautiful. But if everyone does this, the game shifts again. Now you need to pick the faces you think others think others will find beautiful. And so on.

Keynes: “It is not a case of choosing those which, to the best of one’s judgment, are really the prettiest, nor even those which average opinion genuinely thinks the prettiest. We have reached the third degree where we devote our intelligences to anticipating what average opinion expects the average opinion to be.”

This is not a marginal observation about beauty contests. It is a claim about what prices fundamentally represent.

How Deep the Game Goes

Consider a stock trading at £100.

First-degree thinking asks: what is this company worth? You analyse cash flows, competitive position, growth prospects. You conclude it is worth £80. You might short it.

Second-degree thinking asks: what do other investors think it is worth? Perhaps sentiment is bullish. Flows are coming in. Analysts are raising targets. Even if you think it is overvalued, you recognise that others disagree, and their buying will move the price.

Third-degree thinking asks: what do other investors think other investors think? Perhaps everyone privately believes the stock is overvalued. But everyone also believes that everyone else is still buying. So everyone keeps buying, waiting for others to stop first.

This regress has no natural stopping point. Each level of reasoning produces a different answer, and acting on one level while others act on another creates opportunity — or ruin.

The fundamental investor who stops at first-degree thinking treats the market as a weighing machine. Eventually, they believe, price converges to value. They may be right — eventually. But as the old market adage (often attributed to Keynes) goes, the market can remain irrational longer than you can remain solvent. Solvency is a path-dependent property, not an ensemble average. The ergodicity problem from The Path The Maths Misses applies: being right on average means nothing if this particular path bankrupts you.

Common Knowledge and Coordination

Game theorists have a precise concept for what separates private knowledge from the kind that moves markets: common knowledge. Robert Aumann’s foundational work established that something is common knowledge when everyone knows it, everyone knows that everyone knows it, and so on to infinity. This is a much stronger condition than widespread knowledge.

The distinction matters enormously. Many sophisticated investors may privately believe an asset is mispriced — yet each cannot be certain that everyone else will act. Perhaps being early is worse than being wrong. The knowledge is widespread but it may not be common knowledge in the technical sense.

Stephen Morris and Hyun Song Shin pioneered the application of “global games” theory to analyse exactly these situations. One result from their programme is particularly relevant here: when agents face uncertainty not just about fundamentals but about what others believe about fundamentals, small changes in public information can trigger discontinuous shifts in behaviour.

Mutual knowledge, however widespread, is often insufficient to coordinate action. What breaks the stalemate is a signal known to be seen by everyone — one that shifts higher-order beliefs past the threshold where waiting becomes riskier than acting. The gap between “most people know” and “everyone knows that everyone knows” is where coordination failures live.

What Creates Common Knowledge

If coordination problems persist because common knowledge is absent, what makes it suddenly appear? Public signals — not just seen by everyone, but seen to be seen by everyone.

The fairy tale of the emperor’s new clothes is the canonical illustration. Every citizen privately sees the emperor is naked, but each assumes others believe the clothes are real — a state game theorists call mutual knowledge. Then a child shouts the obvious. The content is not new. What is new is that everyone now knows that everyone heard it. Mutual knowledge becomes common knowledge. The equilibrium collapses.

Markets have analogous moments. The visibly failed sale in Haarlem was this kind of signal — not new information about tulips, but a public revelation that buyers had stopped believing other buyers would appear. Morris and Shin show this formally: in games with strategic complementarities, public signals have disproportionate impact precisely because they are commonly observed.

This connects to the dynamical systems thinking from What Makes Stable Things Break. The consensus attractor can appear stable — private doubts accumulate, but the basin holds because no one knows that everyone else doubts too. The common knowledge event is the bifurcation trigger.

Beliefs That Change Reality

So far, the coordination problem has been about beliefs determining prices. Reflexivity is the stronger claim: that prices can then alter the fundamentals themselves.

George Soros formalised this as a dynamic framework. The beauty contest describes a static game: anticipate beliefs, trade accordingly. Reflexivity adds a feedback loop: trading changes the reality that beliefs are about. The object of the beauty contest is not fixed, because the contestants are changing it as they play.

Consider a bank whose solvency depends on confidence. If investors believe it is solvent, they lend and deposit; the bank remains liquid and therefore solvent. If they believe it is insolvent, they withdraw; it becomes illiquid and therefore insolvent. The belief creates the reality it describes — a genuine feature of systems where prices are inputs to the system that generates them, not merely estimates of some external reality.

Reflexivity means that “fundamental value” is not always a fixed star by which prices can be judged. Value itself can depend on the path of prices, because prices affect the fundamentals. Reflexivity is strongest where prices affect solvency constraints: banks, collateralised lending, pegged currency regimes. It is weaker in mature large-cap equities where the feedback from price to fundamentals is slower and more diffuse.

The Coordination Question

None of this means fundamentals are irrelevant. Cash flows arrive. Earnings are reported. Over sufficiently long horizons, prices must accommodate these realities. But fundamental analysis answers one question — what is this asset worth? — while leaving another unanswered: when will prices converge to this value, and what path will they take? These are questions about belief dynamics and coordination. They require a different model — not because fundamentals are wrong, but because fundamentals are incomplete.

The investor who knows a stock is mispriced has solved one problem. They have not solved the problem of whether the mispricing will correct before they are forced out of their position. Abreu and Brunnermeier formalised this as synchronisation risk: rational arbitrageurs recognise the mispricing at different times, so each is uncertain who else has figured it out and when they will act. Being early is costly, so everyone waits — a coordination failure among the very people who should be restoring efficiency.

Several things follow. Catalysts matter as much as mispricings — a mispricing without a catalyst is a statement about static value, while a catalyst is a mechanism for converting private knowledge into common knowledge. The persistence of prices that diverge from fundamentals is not anomalous; it is the natural state of a coordination game where common knowledge has not yet formed. And reflexivity means that your own actions — if large enough — change the game. The observer is inside the system.

The coordination structure extends well beyond markets. In technology adoption, a platform can dominate not because users prefer it, but because users believe other users prefer it — the upstart with a better product faces the same problem as the fundamental investor with the right valuation. Currency works the same way: people accept a £20 note not because the paper is worth anything, but because they believe others will accept it too. The coordination equilibrium holds until it does not — and when it breaks, as in hyperinflation episodes, it breaks fast.

The beauty contest is not a flaw in markets to be corrected. It is the structure of any game where participants must anticipate each other’s actions. Understanding it does not make you immune to its dynamics.

The discipline is in asking: what game are the other players playing?

References & Further Reading

1. The General Theory of Employment, Interest and Money: John Maynard Keynes

2. Global Games: Theory and Applications: Stephen Morris & Hyun Song Shin

3. Agreeing to Disagree: Robert Aumann

4. The Alchemy of Finance: George Soros

5. Social Value of Public Information: Stephen Morris & Hyun Song Shin

6. Bubbles and Crashes: Dilip Abreu & Markus Brunnermeier

7. The Limits of Arbitrage: Andrei Shleifer & Robert Vishny

Here’s a bet. A coin is flipped. Heads, your entire wealth increases by 50%. Tails, it decreases by 40%. The expected value is positive: 0.5 × (+50%) + 0.5 × (-40%) = +5% per flip. A good bet, by the standard criterion.

Now play it repeatedly. Start with £100. After one round, you might have £150 or £60. After two rounds, the possibilities are £225, £90, £90, or £36. Keep going. The result is striking: over time, almost everyone goes broke.

This isn’t bad luck. It is mathematics. The expected value—the average across all possible outcomes—is positive. But you don’t experience the average across possible outcomes. You experience one sequence of outcomes, and in that sequence, the losses compound multiplicatively against you. A 50% gain followed by a 40% loss doesn’t return you to even; it leaves you at 90% of where you started. Do that enough times, and you converge toward zero.

The bet has positive expected value but negative expected growth. And you live in time, not in an ensemble of parallel worlds.

What kind of problem is this?

When something looks rational by one criterion but leads to ruin by another, that is a specific shape of problem. The math isn’t wrong. The framework is incomplete—it is answering a question different from the one that matters.

Expected value answers the question: “If I could play this game across many parallel versions of myself simultaneously, what would the average outcome be?” That is meaningful for casinos, insurance companies, and anyone who can diversify across thousands of independent bets. But it is not the question facing someone with one life, one portfolio, and one sequence of returns compounding through time.

The question that matters for a mortal is: “What happens to me, over time, playing this game repeatedly?”

The discipline that has thought rigorously about when these two questions give the same answer—and when they don’t—is statistical mechanics. And more recently, a physicist named Ole Peters has formalized the distinction with uncomfortable precision.

So let’s borrow.

This is the multi-model move: recognize the shape of a problem, find the discipline that has thought rigorously about that shape, and import its frameworks deliberately rather than reinventing them from scratch.

The standard framework was asking a different question. “What is the expected value?” is not the same as “What happens over time?” These questions coincide only under special conditions—and wealth, which compounds multiplicatively, doesn’t meet those conditions.

Ergodicity and Its Absence

In physics, a system is “ergodic” if its time average equals its ensemble average. Imagine measuring the speed of a gas molecule. You could track one molecule over a long period and average its speeds through time. Or you could measure all molecules at one instant and average across the population. For an ergodic system, these two averages converge. The one-over-time and the many-at-once give the same answer.

Much of standard financial reasoning implicitly assumes ergodicity—treating expected values as decision-relevant for individuals, which assumes the average across possible states equals the average through time. For additive processes, this works: if I gain or lose fixed pound amounts, my long-run average equals the expected value.

But wealth isn’t additive. It is multiplicative. Returns compound. A 50% loss requires a 100% gain to recover. A sequence of +20%, -20%, +20%, -20% doesn’t leave you flat; it leaves you at 92% of where you started. The order and compounding change everything.

For multiplicative processes, ensemble averages and time averages can diverge dramatically. The ensemble average—what happens to the average across many simultaneous players—can grow, even as the time average—what happens to a single player over many rounds—shrinks. Most players go broke even as the “average” player gets richer. A vanishingly small fraction of paths—those lucky enough to string together heads after heads—accumulate astronomical wealth, pulling the average up even as the typical player converges toward zero.

This is the ergodicity problem. The expected value is real, but you can’t access it. You are stuck in time, experiencing one path through the possibility space, and that path has different statistics than the ensemble.

The Information Connection

Boltzmann defined entropy in 1870s thermodynamics as

where W is the number of accessible microstates—the count of ways the system could be arranged. Shannon defined entropy in 1948 information theory as

and he explicitly modeled it on Boltzmann. When all states are equally likely, Shannon’s formula reduces to H = log W—the same structure, different units.

Jaynes completed the circle in 1957, showing that statistical mechanics itself derives from maximum entropy inference. They aren’t parallel theories; they are one theory wearing different clothes.

So when we say a system is non-ergodic, we are making an information-theoretic statement: the ensemble contains states your path will never visit. The expected value aggregates outcomes you will never experience. The ensemble gives you information about a space you don’t traverse.

That information gap is the risk. Risk isn’t uncertainty. Risk is uncertainty resolving faster than you can adapt.

The Kelly Criterion

If expected value is the wrong objective, what’s the right one? For multiplicative processes, the answer is: maximize the expected growth rate—the time-average rate at which wealth compounds.

This leads to the Kelly criterion, derived by physicist John Kelly in 1956—mathematically equivalent to the solution Daniel Bernoulli proposed for the St. Petersburg paradox in 1738: maximize the logarithm of wealth, not the raw pound amount.

The insight: optimal bet sizing for time-average growth is always smaller than what expected-value maximization would suggest. Often much smaller. The Kelly criterion is conservative relative to expected value precisely because it accounts for the compounding of losses.

A full Kelly bet is often considered aggressive in practice—many practitioners use “half Kelly” or less, because the criterion assumes you know your true edge with certainty, and overestimating your edge is ruinous. But even full Kelly is far more conservative than “bet as much as possible on positive expected value.”

The Kelly framework makes explicit what the expected-value framework obscures: survival constraints bind. You can’t compound wealth if you have been wiped out. The path matters, not just the destination.

Volatility Drag

The ergodicity problem shows up in places that don’t look like coin flips. Consider leveraged ETFs—funds that promise 2x or 3x the daily return of an index.

Suppose an index goes up 10% one day and down 10% the next. It ends at 99% of where it started (1.10 × 0.90 = 0.99). A 2x leveraged ETF goes up 20% and down 20%: it ends at 96% (1.20 × 0.80 = 0.96). A 3x leveraged ETF: up 30%, down 30%, ending at 91% (1.30 × 0.70 = 0.91).

The index ended down 1%—roughly flat—but the leveraged products lost far more. This is “volatility drag”—the systematic decay that affects any leveraged position in a volatile asset, even if the underlying goes nowhere. It is a mathematical consequence of multiplicative compounding, not a fee.

Volatility drag is the ergodicity problem in disguise. The expected daily return of a 2x leveraged ETF is 2x the expected return of the index. But the time-average return—what you actually experience over a sequence of days—can be negative even when the index’s arithmetic average daily return is zero. The ensemble average and time average diverge.

This catches people because the ensemble framing is intuitive. “2x the return” sounds like more is better. But you experience one path, and on that path, the volatility compounds against you.

The Exception

There is a reason expected-value thinking persists in finance: under certain conditions, it is approximately correct.

An investor can narrow the gap between ensemble and time averages by running many truly independent bets simultaneously and staying in the game long enough for the law of large numbers to work.

What does meeting these conditions require?

The bets must be truly independent—not just different asset names, but genuinely uncorrelated return drivers. Correlation spikes in crises, counterparty concentration, crowded positioning, and hidden beta in private assets can all collapse a thousand “independent” positions into one. Diversification only counts if the bets stay independent when it matters.

There must be sufficient scale: the law of large numbers needs numbers, and a handful of positions isn’t enough. You need enough independent bets that the variance of your realized path shrinks toward the ensemble mean.

You need buffers to absorb shocks—capital, liquidity, time—that determine how much uncertainty you can absorb before being forced to act. Leverage, redemptions, and margin calls can all force you off the path before independence can save you. And you need the ability to rebalance: winners must subsidize losers through a shared balance sheet. If you can’t move capital between bets, each bet eventually goes to zero on its own. The ensemble magic requires pooling.

Who has the structure to meet these conditions? A large, well-capitalized insurer running thousands of uncorrelated policies. A sovereign wealth fund with a multi-decade horizon and no redemption pressure. A market maker capturing small edges across millions of trades. These are the structural capacities that must be actively maintained. Most investors don’t meet all four. They are closer to the single path than they think.

Long-Term Capital Management had two Nobel laureates, diversification across global bond and derivatives markets, and 25:1 leverage. In 1998, when Russia devalued and imposed a debt moratorium, correlations spiked and their “independent” bets moved together. The leverage meant they couldn’t survive long enough for independence to reassert itself. The ensemble average was irrelevant; they experienced one path.

These conditions tend to fail together. The investor who looked ergodic in calm markets discovers in the crisis that they were always on one path—they just couldn’t see it until the path turned.

The Time-Average Question

When does this lens apply? Whenever outcomes compound multiplicatively.

This includes most of investing. Returns compound. Losses require larger gains to recover. Sequences matter. The ergodicity problem isn’t an edge case; it is the default condition for many investors managing wealth through time.

But it extends beyond finance. Career decisions compound—an early setback can close off paths in ways that don’t average out. Health decisions compound—some risks, if they materialize, aren’t recoverable. Organizational decisions compound—a bankruptcy isn’t offset by an alternate-universe success. Any domain with irreversibility and path-dependence is a domain where ensemble thinking misleads.

The framework asks: Am I thinking in terms of what would happen on average across parallel instances? Or what will happen to me, sequentially, over time? If the first, and if outcomes compound multiplicatively, I may be answering the wrong question.

The discipline is in asking: not what will happen on average, but what will happen to me?

References & Further Reading

1. The Time Resolution of the St. Petersburg Paradox: Ole Peters

2. Ergodicity Economics: Ole Peters & Alexander Adamou

3. Fortune’s Formula: William Poundstone

4. A New Interpretation of Information Rate: John Kelly

5. A Mathematical Theory of Communication: Claude Shannon

6. Information Theory and Statistical Mechanics: E.T. Jaynes

7. Thinking in Bets: Annie Duke

In late July 2024, the yen carry trade was a heavily-positioned source of global funding and leverage. The logic was straightforward: borrow in yen at near-zero rates, invest in higher-yielding assets, collect the spread. The Bank of Japan had been dovish for years. Rate markets priced the differential continuing. Options priced some tail risk, but the base case was continuation. Consensus was clear: the state holds.

On July 31, the BoJ raised its policy rate to 0.25%. Two days later, a weak US jobs report sent US yields plunging—the carry differential was squeezed from both sides. Within days, the trade unwound violently. The Nikkei had its worst single-day drop since 1987. Volatility spiked globally. Positions that had seemed unrelated turned out to be linked through common yen funding. By August 5, the stress was acute. By August 9, much of the move had reversed—but portfolios had been damaged, and the consensus that had seemed so stable had revealed itself as fragile.

What kind of problem is this?

The mechanics of carry were unchanged—you could still borrow yen cheaply and invest elsewhere. But the assumptions that made it feel stable—wide differentials, low FX volatility, a one-way yen narrative—had shifted. What failed was the assumption that the state would persist—that the parameters governing the trade would remain in the range where the strategy made sense. The consensus wasn’t wrong about the current state. It was wrong about the stability of that state.

When systems appear stable but can flip suddenly, when gradual parameter drift precedes discontinuous breaks, when the consensus view is not “X will happen” but “the current state will persist”—that is a specific shape of problem. And one discipline that has thought carefully about state persistence and sudden transitions is dynamical systems theory.

So let’s borrow.

This is the multi-model move: recognize the shape of a problem, find the discipline that has thought rigorously about that shape, and import its frameworks deliberately rather than reinventing them from scratch.

A caveat before we proceed. Dynamical systems concepts—attractors, bifurcations, critical transitions—have a mixed record when applied to financial markets. Studies that look for “critical transitions” in price series often find weak or inconsistent results. The reason: markets are not physical systems, and applying the framework generically misses what makes it useful.

The concepts work when you specify what consensus is actually about. In a carry trade, the attractor isn’t “the market”—it is a specific belief: “the yield differential persists.” Staying in the trade isn’t neutral; it is an active forecast that this belief will hold. The position IS the forecast. There is no “just harvesting premium”—that phrase hides the bet.

This is why the framework applies to carry trades, volatility selling, and currency pegs: in each case, consensus is anchored on state persistence, and holding the position requires the current parameters to remain stable. The question isn’t whether markets show critical transitions in general. It is whether this specific consensus—the one your position depends on—is stable.

Attractors and Basins

Dynamical systems theory studies how systems evolve over time and what states they tend toward.

An attractor is a state the system gravitates toward. Think of a ball in a bowl: disturb it, and it rolls back to the bottom. The basin of attraction is the set of starting points that lead to that attractor—anywhere on the inside of the bowl leads to the same resting place.

In markets, a consensus view can function as an attractor. “The yen stays weak” was not just a belief; it was a state that the system tended toward. Information, capital, and positioning flowed toward it. Analysts who deviated faced career risk. Trades that bet against it lost money, reinforcing the consensus. The attractor pulled.

In markets, the attractor isn’t the belief by itself—it is the observable state the belief sustains: prices, volatility, leverage, and funding conditions reinforcing one another.

But attractors are not permanent. They exist within basins, and basins can shrink. The parameters that define the landscape—in this case, BoJ policy, positioning concentration, the behavior of stabilizing flows—can drift. The ball still sits at the bottom of the bowl, but the bowl itself is getting shallower.

A bifurcation is the point where a small change in parameters causes a qualitative change in behavior. The attractor can weaken, split, or disappear entirely. Before the bifurcation, the system looks stable—disturbances get absorbed, the consensus holds. After the bifurcation, the system behaves completely differently.

The danger is being right about the current attractor while missing that the basin is shrinking. The consensus may be correct about the present state, but wrong about how stable that state is. This is not a forecasting error about what will happen; it is a stability error about whether the current state will persist.

Sensors and Validity Checks

This distinction matters: a validity check tells you the state has broken; a sensor might tell you the state is becoming fragile before it breaks.

In the yen carry unwind, validity checks included VIX, funding spreads, and realized volatility—all of which spiked during the crisis, confirming the break but providing no lead time. By the time they fired, the damage was underway. In this episode, even the VIX spike was partly microstructure-driven—pre-market prints distorted by wide spreads and thin liquidity. A useful break marker, but not a clean signal.

Potential sensors would have tracked different signals. Mean reversion decay: are shocks still being absorbed, or are they persisting longer? Position clustering: how correlated are margin calls across different funds running similar trades? Amplification ratio: is a secondary move larger than the primary shock—a sign that feedback mechanisms are strengthening?

These measure approach to criticality, not arrival at crisis. The lesson is not that crises are predictable—they aren’t, with precision. The lesson is that stability itself can be assessed. The question shifts from “will the state break?” to “how stable is this basin?” Parameters can be tracked. Fragility can be measured, even when timing cannot.

Where Stability Hides Fragility

If the pattern holds, where else might it apply?

Volatility selling has the same structure as the yen carry trade. Some strategies are explicitly short volatility—selling puts, selling variance swaps, shorting VIX products. Others have implicit short-vol properties: vol-targeting funds and risk parity strategies that mechanically delever when volatility spikes. Both collect premium in calm markets; both face the same feedback risk. The consensus is that vol remains contained. The position IS the forecast: staying short vol is betting the low-volatility state persists.

The parallel to carry is precise. Both collect a premium for bearing state-break risk. Both have mechanical feedback: in carry, the unwind forces selling of the funded asset; in short vol, a spike forces delta hedging, which amplifies the move, which forces more hedging. Both can cascade from within—market dynamics, not external shocks.

After each volatility spike—2018’s “Volmageddon,” the March 2020 COVID crash, the August 2024 yen unwind—short vol positioning rebuilds. The trade works again. The premium is there. The consensus re-forms: volatility stays contained.

The question the framework asks is not “will vol spike?” No one knows the timing. The question is: how stable is this basin?

Start with positioning. Has it re-clustered since the last spike? Systematic strategies, insurers, and retail products all run similar exposures through different instruments. The concentration may be hidden—different labels, same trade.

Then consider the stabilizers. When vol rises, who buys the dip? Are those buyers still active, or are they exhausted? Mean reversion in volatility depends on stabilizing flows. If those flows are weaker than before, shocks persist longer—a sign the basin is getting shallower.

Finally, track amplification. When vol moves, how much does the secondary move exceed the primary shock? When dealers are net short gamma, hedging flows can amplify moves well beyond what the underlying shock would suggest. If the amplification ratio is increasing, the feedback mechanism is strengthening.

None of this predicts the timing of a break. But it tracks the stability of the state. The consensus attractor—”vol stays low”—may be correct about the current state while being wrong about how stable that state is. Validity checks will only confirm that after the fact. Sensors might provide earlier warning—if you are watching.

The Stability Question

Dynamical systems concepts are tools, not truths. Markets adapt; physical systems don’t. A visible fragility signal gets traded on, which can either accelerate the break or postpone it by reducing crowding. The observer affects the observed.

The concepts apply when three conditions hold: the system has an equilibrium it tends toward, that equilibrium depends on parameters that could shift, and feedback mechanisms could amplify disturbances. When these conditions hold, the language of attractors and bifurcations describes something real.

When they don’t—when there is no clear equilibrium, when parameters are stable, when feedback is weak—the concepts add little. The tool fits the problem, or it doesn’t.

The question is whether a situation has this structure. Not “what state is the system in?” but “how stable is that state?” Not “will things change?” but “what is the current equilibrium resting on?”

In careers, the stability question applies when your role depends on conditions that feel permanent but aren’t: a sponsor’s support, a strategy’s favor, a skill’s relevance. The attractor is real—until the parameters shift. In organizations, it applies to practices, business models, and market positions that seem entrenched. Stability isn’t permanence; it is a basin that could be shrinking while everything looks fine. In health, chronic stress or deferred maintenance can make a system brittle in ways that don’t show—until a small shock triggers a transition that wouldn’t have happened before.

The yen carry trade looked stable. Low-volatility states look stable. Many equilibria feel like safety—everyone is there, the logic is sound, the position is working.

The discipline is in asking: is this stable—and what is that stability resting on?

References & Further Reading

1. Nonlinear Dynamics and Chaos: Steven Strogatz

2. Critical Transitions in Nature and Society: Marten Scheffer

3. Stabilizing an Unstable Economy: Hyman Minsky

4. Limits of Arbitrage: Shleifer & Vishny

5. Predatory Trading: Markus Brunnermeier & Lasse Pedersen

6. Yen Carry Trade Unwind (August 2024): BIS Bulletin No. 90

In the summer of 2008, risk managers at major financial institutions could show you their correlation matrices. They knew how their assets moved together. They had stress tests, value-at-risk models, scenario analyses. By the standard measures, many portfolios looked diversified.

Then Lehman Brothers filed for bankruptcy, and the financial system seized up in ways that few had anticipated. Assets that had seemed unrelated crashed together. Institutions that had no direct exposure to subprime mortgages found themselves unable to fund their operations. The contagion spread through channels that didn’t appear in any correlation matrix.

What happened?

The standard tools measured co-movement—how asset prices moved together historically. But the crisis didn’t propagate only through co-movement. It propagated through connections—counterparty relationships, funding dependencies, collateral chains—and the feedback dynamics they enabled. The failure of one node didn’t just correlate with failures elsewhere; it caused failures elsewhere, through a web of obligations that had no statistical signature until it was too late.

What kind of problem is this?

When failures cascade through hidden linkages, when the structure of connections determines which shocks stay contained and which ones spread, that’s not only a statistics problem. It’s a propagation-through-structure problem—where the network matters.

And the discipline that has spent decades studying how structure determines contagion is network science.

So let’s borrow.

This is the multi-model move: recognize the shape of a problem, find the discipline that has thought rigorously about that shape, and import its frameworks deliberately rather than reinventing them from scratch.

The standard approach was asking a different question. “How correlated are these assets?” is not the same as “How are these assets connected?” Correlation is a statistical shadow. Connections encode potential spillover channels—sometimes contractual and direct, sometimes indirect through common exposures.

If we look at financial systems as networks, a different picture emerges. Their fragility—their vulnerability to cascading failure—is not a property of any single node. It is a property of the topology: who is connected to whom, how concentrated those connections are, and what happens when a critical node fails. Miss the topology, and yesterday’s diversified portfolio can become exposed in unexpected ways.

Connected, Not Correlated

Here is the core distinction: correlation tells you that two things tend to move together. It says nothing about why they move together, or what would happen if one of them failed.

Two stocks might be correlated because they are in the same sector, exposed to the same economic factors. If one company goes bankrupt, the other is fine—they were correlated but not connected. Alternatively, two stocks might appear uncorrelated in normal times but be linked through a shared counterparty, a common lender, or the same few market makers. In a crisis, when that hidden node fails, both stocks move together—they were connected but not correlated.

Network science offers a vocabulary for this. A network is a set of nodes (entities) and edges (connections between them). The topology of the network—who is connected to whom, how many connections each node has, whether the network has hubs or is evenly distributed—determines how shocks propagate.

Consider two networks with the same number of nodes and edges. In one, connections are distributed roughly evenly—everyone has about the same number of links. In the other, a few nodes are massively connected while most have only one or two links. These networks share the same density and average degree but have completely different failure dynamics. In the first, a random node failure affects only its immediate neighbors. In the second, if a hub fails, cascades can reach the entire network in steps.

The financial system in 2008 looked more like the second type. A handful of institutions and markets—Lehman, AIG, and the short-term funding markets—were hubs through which vast amounts of risk flowed. When those hubs failed or wobbled, the damage radiated outward through connection pathways that correlation matrices couldn’t see.

Here is the challenge for anyone assessing systemic risk: you can estimate correlations from price data, but you usually can’t observe the network directly.

Counterparty exposures are private. Funding relationships are opaque. The web of who-owes-what-to-whom exists in legal contracts and internal ledgers, not in public data. You’re trying to assess systemic risk with a map that shows terrain but not roads.

This creates a specific kind of limitation. Diversification strategies that look robust under correlation analysis can be fragile under network analysis. When many institutions diversify into the same “uncorrelated” assets using the same prime brokers and the same repo counterparties, network concentration can build up that won’t be visible until a stress event reveals it.

The 2008 crisis was a case study in this. Portfolios that looked diversified were linked through common funding sources. Assets that seemed unrelated shared exposure to the same structured products. When the network came under stress, the “uncorrelated” assets moved together because they were all trying to exit through the same doors, sell to the same buyers, and roll the same funding.

The correlation matrix updated fast—by October 2008, cross-asset correlations spiked and diversification benefits collapsed. But that was the symptom, not the cause. The cause was network topology that concentrated risk in ways the statistical tools couldn’t detect.

Contagion

Network science has identified several mechanisms that determine how shocks propagate.

Start with degree distribution. If connections are concentrated in a few hubs, the network is robust to random failures but highly vulnerable to hub failures. This is the “too connected to fail” problem, which is distinct from “too big to fail.” A small node that connects many others can be more systemically important than a large node with few connections.

There is a reason hubs form: efficiency. TSMC became central to AI because it offered the best capabilities at the lowest cost. The pre-2008 interbank market concentrated in a few dealers because they offered the deepest liquidity. Network science reveals the trade-off: systems optimized for efficiency tend to develop hubs, and hubs create fragility to targeted shocks. The same concentration that makes normal operations smooth makes disruptions catastrophic.

Then consider path length. Many financial networks exhibit short path lengths and core-periphery structure—a dense core of major intermediaries with a periphery of smaller players. Contagion spreads fast when any node can reach any other quickly.

Clustering matters too. If your counterparties are also counterparties to each other, you’re in a cluster. Clusters can contain contagion locally, but if a shock breaches the cluster boundary, it can spread to other clusters rapidly.

Finally, even without direct connections, nodes can be linked through common exposures to a third entity. This creates hidden network structure that only becomes visible when the common exposure fails.

The key insight is that the same shock can have completely different effects depending on where in the network it hits and what the network structure looks like. A loss at a peripheral node might be absorbed easily. The same-sized loss at a hub might trigger a cascade. Once leverage and liquidity thresholds are breached, network position can matter as much as—or more than—shock size.

This changes how one might assess risk. Correlation-based diversification is necessary but not sufficient. It is possible to build a portfolio with low pairwise correlations and still have concentrated network exposure if the positions share counterparties, funding sources, or liquidity providers. The question isn’t just “do these assets move together?” It’s “are these assets connected through nodes that could fail?”

Tail risk, in this view, is about network structure, not just return distributions. Fat tails in financial returns can reflect cascade and feedback dynamics—defaults, margin spirals, fire sales—not just bigger shocks. Historical volatility alone can’t capture this; understanding the topology that enables cascades matters too.

And “uncorrelated” is a statement about the past, not the future. Assets that have no statistical relationship in normal times can become highly correlated in stress if they’re connected through a common node that fails. The correlation matrix is a fair-weather friend.

Network science doesn’t give you a crystal ball. The full network usually isn’t observable, and network structure itself changes over time as institutions form new relationships and sever old ones. But the framework changes what to pay attention to. Instead of asking “what’s the correlation between A and B?”, the question becomes: “what connects A to B?” Instead of assuming diversification protects a portfolio, the ask is: “am I diversified across the network, or just across the correlation matrix?”

Where Fragility Hides

If the network lens reveals anything, it is that opacity is where risk accumulates unobserved. In 2008, the hidden topology was in structured products and counterparty chains. Where might similar dynamics exist today?

Consider AI supply chains. A portfolio holding Nvidia, AMD, Microsoft, Google, and Meta looks diversified across the technology sector. Correlation analysis would show these stocks move together—but correlation only tells you that they move together, not why. The network tells you why: trace the physical supply chain, and most roads lead to TSMC.

Most leading-edge AI accelerators—Nvidia’s GPUs, AMD’s accelerators, Google’s custom TPUs—rely on a single foundry ecosystem, with critical capacity still concentrated in Taiwan. The network has a hub that doesn’t appear in any standard risk model. CoWoS packaging capacity has been an even tighter bottleneck; constraints there ripple through every company building AI infrastructure.

This isn’t a financial counterparty network—it is a physical supply chain. But the topology is the same: hidden concentration that creates correlated exposure across positions that look independent. A disruption at the hub—whether from geopolitical tension, natural disaster, or capacity constraint—would affect “diversified” AI positions simultaneously. The correlation matrix would update fast. But that would be the symptom, not the cause.

The network lens doesn’t predict when disruption comes or whether it comes at all. TSMC might continue operating smoothly for decades. What the lens does is identify the structure: where are the hidden hubs that create correlated exposure? In AI infrastructure, the answer is legible if you look at the physical network rather than the financial correlations.

The Network Question

The correlation view and the network view each reveal something the other misses. Correlation tells you how assets have moved together. Networks tell you how they’re connected—and how shocks might propagate.

When does this lens apply? When you are assessing systemic risk, not just asset risk. When failures might cascade rather than stay contained. When the question isn’t “how volatile is this?” but “what is this connected to?”

The framework asks specific questions. Who are you connected to—not correlated with, but actually connected to through contracts, counterparties, or common dependencies? Which of those connections run through hubs—nodes that many others also depend on? And what happens to you if a hub fails, even one you have never heard of?

These questions apply wherever failure propagates through structure rather than just co-occurring statistically.

In supply chains, the risk isn’t that your suppliers are “correlated”—it is that they all depend on the same port, the same chip fab, the same logistics provider. In organizations, fragility isn’t about departments having similar performance—it is about critical knowledge or decisions flowing through a single person. In technology systems, resilience isn’t about redundant servers—it is about whether those servers share a power grid, a cloud provider, or a software dependency.

The network lens asks the same question in each case: where are the hidden hubs?

We used finance to make the lesson concrete, but the pattern is general. Whenever you find yourself saying “I didn’t see how those two things were related”—after a cascading failure reveals the connection—that’s a network problem. The correlation tools couldn’t see it because they measure co-movement, not connection.

The discipline is in asking: what is the network I cannot see?

References & Further Reading

1. Network Science: Albert-László Barabási

2. Complex systems: Ecology for bankers: Robert May, Simon Levin & George Sugihara

3. Systemic risk in banking ecosystems: Andrew G. Haldane & Robert M. May

4. Deciphering the Liquidity and Credit Crunch 2007-2008: Markus Brunnermeier

5. Systemic Risk and Stability in Financial Networks: Daron Acemoglu, Asuman Ozdaglar & Alireza Tahbaz-Salehi

6. Thinking in Systems: A Primer: Donella Meadows

By the late 2010s, value investing faced its longest test.

The strategy had one of the longest track records in finance. Buy cheap stocks, sell expensive ones, wait. Benjamin Graham had codified it in the 1930s. Decades of academic research documented a historical value premium—cheap stocks, as measured by metrics like price-to-book, systematically outperformed expensive ones over long periods. Generations of investors built careers on it.

Then it entered a decade-plus drawdown. From mid-2007 to late-2020, systematic value—rules-based strategies that buy cheap stocks and sell expensive ones—underperformed growth by a significant margin, whether measured by long-short factors or value-vs-growth indices. Disciplined managers who had built decades of credibility watched their core strategy lag year after year. Some clients lost patience. Some funds closed. Many questioned whether value investing was dead.

What kind of problem is this?

When something works and then stops working—without any internal flaw—that is a specific shape of problem. And the shape tells you where to look for insight.

If the machinery had broken, you would call an engineer. If the logic had a flaw, you would call a mathematician. But when something succeeds in one context and fails in another, without itself changing, that is not an engineering problem or a logic problem. It is an adaptation problem. The thing is fine; the fit between the thing and its environment has changed.

And the discipline that has spent 150 years thinking rigorously about the fit between organisms and environments is evolutionary biology.

So let’s borrow.

This is the multi-model move: recognize the shape of a problem, find the discipline that has thought rigorously about that shape, and import its frameworks deliberately rather than reinventing them from scratch.

Both sides of the value debate were asking the wrong question. “Does value work?” is like asking “Is a polar bear fit?” The answer is: fit for what? A polar bear is exquisitely adapted to the Arctic—the apex predator of a frozen world. Drop it in the Sahara and every adaptation becomes a liability.

If we view investment strategies as organisms, their “fitness”—their ability to generate returns—is not an intrinsic property. It is a relationship between the strategy and the environment in which it operates. Change the environment, and yesterday’s perfect adaptation becomes tomorrow’s mismatch.

What the Genome Knows

Evolutionary biology learned this lesson the hard way. Fitness only makes sense relative to an environment—there is no universal fitness ranking.

Christoph Adami, the physicist we met in our previous post, formalizes this precisely. An organism’s genome is essentially a compressed encoding of information about its environment. The polar bear’s white fur, fat reserves, and hunting behaviors are a physical record of what worked in Arctic conditions over thousands of generations. The genome “knows” about the environment, in the sense that it carries information useful for surviving there.

This framing transfers directly to investing. A strategy’s design—its signals, its rules, its parameters—is a compressed encoding of information about the market environment in which it was developed. A value strategy “knows” that cheap assets tend to revert to fair value. But what happens when what the strategy “knows” stops being true?

This is the backtest limitation. Almost every investment strategy that gets deployed passed a backtest—it found patterns in historical data that would have made money. But a backtest is a fitness measurement in a historical environment. It tells you this organism would have thrived in the Pleistocene. It says nothing about whether the Pleistocene still exists.

Value investing’s track record was built in an environment where capital was scarce, information traveled slowly, and investors often overreacted to short-term bad news. Patient investors who bought beaten-down stocks were compensated for providing liquidity and bearing uncertainty. The strategy’s “genome” encoded this environmental structure.

Then the environment shifted. In the US and other developed markets, the cost of capital fell for decades—interest rates dropped from double digits to near zero, making growth more valuable as distant cash flows were discounted less heavily. Information became instantaneous. Systematic strategies proliferated, changing the dynamics of how patterns got traded. The strategy’s genome still encoded the old environment—but that environment had changed.

This isn’t a story about value investing being “wrong.” It is a story about environmental mismatch. The polar bear’s genome is a masterpiece of evolutionary engineering. It just doesn’t help when the ice melts.

Adaptive Lag

Biologists have a concept called “adaptive lag”—the delay between an environmental change and a species’ adaptation to it. If the environment shifts faster than the generation time allows, the species falls out of sync with its world.

Strategies face the same problem with an additional twist: a deployed ruleset doesn’t evolve on its own. A species, given enough time, will adapt through selection. But a strategy sitting in production encodes a snapshot of an environment that may no longer exist.

This is why strategies decay. Not because they were badly designed, but because they were designed for a world that has moved on. The signals that once carried information about future returns now carry noise. The relationships that once held have broken or been arbitraged away.

But there is a deeper problem: in markets, the environment fights back. In biology, organisms adapt to their environment, but the environment doesn’t usually adapt back. Markets are different. When a strategy succeeds, capital flows in. More capital chasing the same patterns can change the patterns themselves. This is reflexivity with an evolutionary flavor: the organism’s fitness changes the fitness landscape.

For some strategies, this dynamic is the dominant story—statistical arbitrage that gets crowded until the spread disappears, or momentum trades that reverse sharply when positioning becomes concentrated. But value’s decade-plus drawdown doesn’t fit neatly into the “arbitraged away” narrative. During this period, value-growth spreads actually widened—cheap stocks got cheaper relative to expensive ones. Something else was happening.

Cheap can get cheaper for rational reasons: if discount rates fall and stay low, growth stocks should command higher multiples. The environment shifted in ways that may have rationally repriced the value-growth gap. The environmental thesis got a partial test in 2022: when rates rose sharply, value outperformed growth—consistent with the framework’s emphasis on environmental fit.

Price-to-book—the canonical value signal—likely became noisier as the economy shifted from tangible to intangible assets. Book value captures factories and inventory well; it captures software and network effects poorly. Many systematic value strategies now use composites (earnings yield, cash-flow yield, sales-based measures) partly to reduce reliance on any single metric. The genome is adapting.

If fitness is relational, then evaluating a strategy requires evaluating the environment—and your beliefs about whether that environment will persist. The strategy’s historical returns tell you it was fit for the past. Your job is to figure out whether the future will look like the past—and to notice when it doesn’t.

Fit for What?

If the evolutionary lens reveals anything, it is the importance of asking what environmental conditions your strategy assumes—and whether those conditions still hold.

Consider the mirror image of value’s long winter: the extraordinary outperformance of the mega-cap technology companies—the hyperscalers. While value languished, a handful of companies delivered returns that dominated global indices. Market-cap weighting mechanically increased exposure to these winners as they rose, amplifying concentration—though whether passive flows drove returns or merely reflected them is debated. For over a decade, “own the mega-cap growers” was the winning strategy.

This wasn’t irrational. These are genuinely exceptional businesses: dominant market positions, network effects, recurring revenues, massive cash generation, and the infrastructure buildout for artificial intelligence. Unlike the dot-com era, these companies sit on cash piles that earn billions in a high-rate world. The fundamentals were real. But the magnitude of outperformance—the degree to which these companies pulled away from everything else—also had environmental roots.

Start with rates. Long-duration assets benefit disproportionately from falling discount rates. When rates fall, the present value of cash flows far in the future rises more than the present value of near-term cash flows. Growth stocks are long-duration assets; their value depends heavily on earnings years or decades away. Four decades of falling rates were a structural tailwind for this kind of company.

Then consider growth scarcity. In a world of sluggish nominal GDP growth, companies that could reliably grow revenues at 15-20% per year became precious. The market paid up for growth because growth was rare. The hyperscalers delivered it consistently, and scarcity justified premium multiples.

Platform economics reinforced the dynamic. Winner-take-all dynamics—network effects, switching costs, data moats—meant that early leaders extended their leads. The market rewarded concentration because concentration reflected real competitive advantages. Leaders extended their leads, and the market recognized that dynamic.

Most recently, the buildout of AI infrastructure created a new growth narrative. The hyperscalers are major financiers and beneficiaries of the AI boom—the cloud platforms, the chip buyers, the foundation model builders. Capital expenditures that would sink most companies became signals of future dominance for these.

The question is whether this environmental fit continues—or whether some of these tailwinds are shifting.

Rates may stay higher—if the four-decade bond bull market is over, the mechanical bid for long-duration assets weakens. AI capex eventually needs returns—if the payoff is slower or more diffuse than priced, the narrative could shift. Growth scarcity could end—if nominal growth picks up, the premium for reliable growers may compress. And regulatory pressure remains an environmental variable that could alter competitive dynamics.

None of this means the hyperscalers are “overvalued” or due for a fall. They might continue to outperform for years. The businesses are real, the moats are deep, and the AI wave may be as transformative as the market believes. But the evolutionary lens asks a specific question: how much of the outperformance was the organism, and how much was the environment?

Any sustained outperformance reflects a set of environmental conditions: falling rates, growth scarcity, platform dominance, and now AI capex. The framework doesn’t tell you to sell. It tells you to ask: which of these conditions are expected to persist, and what happens if that expectation is wrong?

The Adaptation Question

The machine metaphor dominates in many domains. Investment strategies are “built.” Business models are “engineered.” Organizational processes are “optimized.” The language assumes these things are robust and context-independent—that good design transcends circumstance.

The organism metaphor is more honest. Strategies are adapted, not designed from first principles. They carry information about an environment, not universal laws. They can thrive or struggle depending on conditions they don’t control.

When does this lens apply? When something that worked stops working—without any internal flaw. When the machinery is fine but the results have changed. When the debate is “does X work?” rather than “is X fit for this environment?” That’s an adaptation problem, and evolutionary biology has the tools.

The framework doesn’t tell you what to do. It tells you what to ask.

The framework asks three questions. What does your approach “know”—what assumptions are encoded in its design? What context does it assume? And does that context still exist? These questions apply whether you are evaluating a portfolio, a business model, or an institutional process.

The discipline is in asking: fit for what?

References & Further Reading

1. What is Information?: Christoph Adami

2. The Origins of Order: Self-Organization and Selection in Evolution: Stuart Kauffman

3. Adaptive Markets: Financial Evolution at the Speed of Thought: Andrew Lo

4. Is (Systematic) Value Investing Dead?: Israel, Laursen & Richardson

5. Darwin’s Dangerous Idea: Daniel Dennett

Here is a question that sounds simple: how much uncertainty does a coin flip contain?

The standard answer is one bit. Heads or tails, fifty-fifty, log₂(2) = 1. This is the textbook definition of entropy, the information-theoretic measure of uncertainty—assuming a fair coin, and that “outcome” means only heads or tails. It is clean, mathematical, and wrong—or rather, incomplete in a way that matters enormously once you try to apply it to anything real.

Entropy depends on what you are measuring. If you only care about heads or tails, the coin has one bit of entropy. But what if you also care about the angle the coin makes with magnetic north when it lands? Divide the compass into four quadrants, and you have eight possible outcomes (heads-north, heads-east, heads-south, heads-west, tails-north...). If those eight outcomes are roughly equally likely, the entropy is log₂(8) = 3 bits.

The physical coin hasn’t changed. The entropy has tripled. We didn’t discover more uncertainty in the coin; we changed the description—the outcome space we are using to measure it. What changed is the question being asked.

This is the physicist Christoph Adami’s crucial point: entropy is not a property of the object. It is a property of the relationship between an observer and an object, mediated by what the observer is trying to predict and what they already know. Change the observer’s question, change the entropy. Change what the observer already knows, change the entropy again.

Now consider “edge” in investing—the claimed ability to predict returns better than the market. It is often discussed as if it were a substance, something you have or don’t have, like a secret or a skill. But if Adami is right, this framing is confused from the start.

What kind of problem is this?

These debates about edge, efficiency, alpha—they treat these as intrinsic properties. “Does this manager have edge?” “Is the market efficient?” The debates never resolve, and the reason they never resolve is that the terms are underspecified. They are asking about the coin without specifying the measurement.

That is not a disagreement problem. It is a precision problem. And the discipline that has spent decades developing precise language for knowledge, uncertainty, and observation is information theory.

So let’s borrow.

This is the multi-model move: recognize the shape of a problem, find the discipline that has thought rigorously about that shape, and import its frameworks deliberately rather than reinventing them from scratch.

The edge debate has been asking the wrong question. “Does this manager have edge?” is like asking “Does this coin have entropy?” The answer is: edge about what? Measured by what signals? Conditional on what baseline? A manager might have edge about one target and none about another. The question isn’t yes or no. It’s relative to what.

Adami’s framework makes this precise. Any statement about information requires three specifications: the target (what you are trying to predict), the observer (whose uncertainty we are measuring), and the conditioning set (what you already know). Without all three, “information” is meaningless—and so is “edge.”

Let’s make that concrete.

What “Edge” Actually Means

Edge is not a property of the investor. It is a property of the relationship between the investor’s signals and the future, conditional on what is already priced.

This requires three specifications:

The target. What are you trying to predict? The direction of the S&P 500 over the next month? The relative performance of two stocks? The timing of a volatility spike? These are different targets with different entropy profiles. A signal that is informative about one may be useless for another.

The observer. Whose uncertainty are we measuring? Your signals, your models, your information set. Two investors looking at the same market face different uncertainties because they are equipped with different measurement apparatus.

The baseline. What do you already know—or more precisely, what does the market already know? Information theory measures the reduction in uncertainty from a signal, relative to some prior state. If the market has already incorporated a piece of news into prices, that news carries zero additional information for trading purposes, no matter how true or important it is.

Edge, then, is conditional mutual information: the information your signals carry about your target, given what is already reflected in prices. It is not “being right.” It is being right about something the market doesn’t already know, on a question that actually determines your returns.

The Measurement Problem

This reframes the usual debates about market efficiency. The Efficient Market Hypothesis is often stated as “you can’t beat the market” or “prices reflect all available information.” But information theory shows these formulations are imprecise. Information about what? Available to whom? Reflected how completely?

A market can be highly efficient for one target and inefficient for another. The S&P 500’s level tomorrow is probably hard to predict—too many sophisticated observers have already incorporated their views. But the relative value of two small-cap industrial companies with limited analyst coverage? The timing of a volatility regime shift that most participants aren’t even trying to forecast? These are different questions with different entropy profiles.

The physicist’s way of saying this: the market has no single “temperature.” Different observers, measuring different things, experience different levels of noise. What looks like perfect efficiency from one vantage point may look like exploitable structure from another—not because one observer is smarter, but because they are asking a different question.

This explains why the alpha debate never resolves. Value investors point to decades of factor returns and say edge exists. Efficient-market theorists point to the difficulty of beating benchmarks after fees and say it doesn’t. They may both be pointing at real effects—but interpreting those effects depends on the target and on your model of risk premia. The value factor may carry information about long-horizon returns that short-term price movements don’t reflect; or it may be compensation for bearing certain risks. The day-to-day direction of the market may be genuinely unpredictable. These aren’t necessarily contradictory findings; they are measurements of different things.

Conditioning Is Everything

The deepest implication is about the baseline. In information theory, the value of a signal depends entirely on what you are conditioning on—what you already know.

Consider two analysts who both predict that a company will beat earnings estimates. Analyst A made the prediction based on channel checks, supply chain data, and proprietary surveys. Analyst B made the same prediction by reading the company’s own guidance more carefully than most. Both are “right” if the company beats. But their edge profiles are completely different.

If the market has already incorporated the guidance—if prices reflect the information Analyst B used—then B has zero edge despite being correct. The information was already in the conditioning set. Analyst A, using signals the market hasn’t processed, may have genuine conditional information.

This is why “being right” is insufficient. The relevant question is never “do I know something true?” It’s “do I know something true that isn’t already priced, about a target that determines my returns, on a horizon where it matters?”

The formula for edge isn’t knowledge. It’s conditional knowledge.

The Target Isn’t Reality

But there’s a further subtlety. What exactly is the target you are trying to predict?

The naive answer is “reality”—future earnings, future growth, whether the technology works. But this misses something important. Markets don’t price reality directly. They price a kind of aggregate belief state—beliefs about reality, weighted by capital and risk tolerance, shaped by constraints. This isn’t the same as what people “think” in some survey sense. It is what gets expressed when money moves under real-world frictions: leverage limits, funding conditions, regulatory constraints, mandate restrictions. The pricing state includes not just what investors believe, but what they are able and willing to do given their constraints.

This means the true target is future consensus: the belief-and-constraint state the market will hold at your horizon, as reflected in prices. Edge is information about where that state is going, conditional on where it is now.

An investor who correctly forecasts earnings but can’t say how or when the market will reprice that information has a weaker claim to edge than one who understands the dynamics of belief revision—even if the latter’s fundamental views are less sophisticated. And an investor who ignores constraints—who says “the fundamentals are obvious, the market must reprice”—is missing half the mechanism. Markets can “know” something and still not price it, because constraints prevent the capital from flowing.

This explains a common frustration: “I was right, but I lost money.” You were right about reality. But you weren’t right about consensus. The market didn’t move to your view on your timeline, and in the interim, prices went against you. Being right about fundamentals is necessary but not sufficient. You need a view on fundamentals and a view on how the belief-and-constraint state will evolve.

At long horizons, fundamentals matter more—reality has more opportunities to force belief updating. Earnings arrive, defaults occur, technologies either work or don’t. But even then, you are not escaping consensus; you are betting that consensus will be forced to converge toward reality. That’s still a prediction about beliefs, not just about the world.

What follows from all this? Two things. Evaluating your own edge requires specifying all three components: target, observer, baseline. Vague claims like “I understand this company better than the market” are meaningless until you say what you are predicting, what signals you are using, and what you think is already priced. And edge is not conserved across questions—a signal that is highly informative for one target may be noise for another. It is common to transfer confidence across domains, assuming that insight into one question implies insight into related questions. Information theory says no. Each target has its own entropy, its own conditioning set, its own edge calculation.

The Observer’s Toolkit

Adami’s framework suggests a practical discipline. Before claiming to know anything predictively, you need to answer three prior questions—and these questions often go unasked.

The first is: what exactly am I trying to predict? Not “things will go well” but something precise enough to be wrong about. In investing, that might mean “the stock will outperform its sector by more than 5% over the next six months.” In medicine, “this treatment will reduce symptoms within two weeks.” In strategy, “this initiative will increase retention by Q3.” Precision forces you to confront the actual target. Vague predictions can’t be evaluated, which is why they are so popular.

The second is: what do I know that could inform this prediction? This means listing the signals, being specific about what you are actually measuring. In investing: channel checks, model outputs, management read, macro view. In medicine: test results, patient history, clinical markers. In any domain: what are your measurements? If you can’t name them, you are not reasoning; you are vibing.

The third is the hard one: what is already known? In investing, this means what’s priced. In other domains, it means the baseline expectation—what would a well-informed observer already predict? The value of your signal depends entirely on what it adds beyond that baseline. Lots of information is technically available but not yet incorporated into the consensus view, because attention is limited and processing is costly. That gap is where genuine predictive advantage lives.

If you can’t answer these three questions, you don’t know whether you have information. You have a feeling, a thesis, a view. These are not the same thing.

The coin flip looks simple. One bit of entropy, everyone agrees. But the moment you try to predict something real—a market, a diagnosis, a strategy’s success—the entropy depends on who is asking, what they are asking, and what they already know. Information is in the eye of the beholder.

The discipline is in asking: which eye, looking at what, through what measuring device?

References & Further Reading

1. What is Information?: Christoph Adami

2. The Model Thinker: What You Need to Know to Make Data Work for You: Scott E. Page

3. A Mathematical Theory of Communication: Claude Shannon

4. The Efficient Market Hypothesis and Its Critics: Burton Malkiel

In the second week of August 2007, some of the most sophisticated quantitative hedge funds in the world suffered drawdowns that erased months or years of gains in a matter of days. Major quant funds, down nearly a quarter of their value in weeks. Leading quantitative shops—firms staffed with physicists and mathematicians, running strategies backtested across decades of data—all losing capital simultaneously.

The models weren’t broken. That was the bewildering part. The statistical relationships these funds exploited—value, momentum, mean reversion—hadn’t suddenly stopped working in any fundamental sense. The backtests still looked good. The math still checked out. What the models couldn’t see was each other.

Each shop had plenty of models—risk, execution, factor attribution—but they shared the same missing model: how crowded, reflexive systems behave when everyone’s edge is the same.

The quant quake, as it came to be called, was a crowding crisis. Too many funds had converged on similar strategies, holding similar positions, operating on similar timeframes. When one fund hit trouble and began liquidating, it pushed prices against every other fund running the same playbook. Their selling triggered more selling. The feedback loop was severe precisely because everyone’s “independent” models had led them to the same trades.

These weren’t amateurs fooled by complexity. They were experts fooled by the boundaries of their models. They had models for statistical arbitrage, for factor returns, for transaction costs and execution. What they lacked was a model for the system they had collectively become part of—a model that would have required ecology, or network theory, or an understanding of how intelligent agents pursuing identical strategies in a shared environment inevitably compete away the very returns they are chasing.

This is the disease of the one true model. And it afflicts brilliant people more than average ones, because brilliant people are more likely to have found a framework that works well enough, often enough, that they mistake it for complete.

Why Investing?

This newsletter is about multi-model thinking—the cognitive practice of holding multiple frameworks simultaneously and letting them interfere with each other productively. That is the actual subject. Investing is simply the backdrop.

But it is a useful backdrop, for a few reasons.

First, investing is complex in the technical sense. Asset prices sit at the intersection of human psychology, institutional structure, information dynamics, game theory, and economic fundamentals. No single discipline owns the phenomenon. This makes it a natural laboratory for multi-model thinking, because single models so reliably fail.

Second, investing provides feedback. Many domains let you hold wrong beliefs indefinitely—you can have a bad theory of history or art and never be conclusively refuted. Markets are less forgiving. Prices move. Predictions resolve. You can be wrong in ways that are unambiguous and costly. This feedback is humbling in a way that is useful for learning.

Third, investing is high-stakes enough to attract rigor. The quants who stumbled in 2007 weren’t casual thinkers. They were deploying the best statistical tools available, with real money on the line. That they still failed—that sophistication within a framework couldn’t save them—makes the case for multi-model thinking more vivid than it would be in a domain where the participants are less formidable.

So: we will use investing as the thread. But the skill being developed is general. It applies wherever you face complex systems that don't respect the boundaries of a single discipline. Investing just happens to be a domain where the consequences of single-model thinking are visible, measurable, and occasionally severe.

The Diversity Prediction Theorem

Scott Page, a complexity scientist at the University of Michigan, proved something remarkable about collective prediction. His Diversity Prediction Theorem can be stated simply:

Collective Error = Average Individual Error − Diversity of Predictions

(This applies technically to squared error, but the conceptual logic holds: diversity provides a mathematical credit against collective inaccuracy.)

Read that again. The accuracy of a crowd’s prediction depends on two factors: how good the individual predictions are, and how different they are from each other. Diversity isn’t just pleasant or fair—it is mathematically essential. A crowd of highly accurate but identical predictors will be no more accurate than any individual member. A crowd of moderately accurate but diverse predictors can be extraordinarily accurate, because their errors cancel rather than compound.

Page developed this theorem studying groups—how do you assemble a forecasting team, a jury, a committee? But the insight applies just as powerfully to the individual mind. Your mental toolkit is a crowd of one. And if that crowd consists of a single model applied with increasing sophistication, you are leaving accuracy on the table that no amount of refinement can recover.

The quant funds of 2007 were individually brilliant. But collectively, they were a crowd with no diversity. Their errors didn’t cancel; they compounded. Each fund’s model was a vote for the same positions, the same timing, the same vulnerabilities. Page’s theorem explains why individually smart people can be catastrophically wrong together.

Why Smart People Get Trapped

There is a seductive logic to mastering a single framework. Depth feels more rigorous than breadth. Expertise is legible, respected, defensible. If you are the person who really understands statistical arbitrage, or constitutional law, or evolutionary psychology, you have an identity, a brand, a claim to authority. The generalist, by contrast, seems dilettantish—a little bit of everything, mastery of nothing.

But this is a confusion of social reward with epistemic accuracy. The expert’s depth is genuinely valuable—within the domain where that expertise applies. The problem is that complex systems don’t respect domain boundaries. A stock price is simultaneously a statement about cash flows, about investor psychology, about liquidity conditions, about narrative momentum, about the positioning of everyone else trying to profit from the same insight. No single discipline owns the phenomenon.

The single-model expert has purchased depth at the price of blindness. And in complex systems, blindness is not merely ignorance of detail. It is vulnerability to the forces you cannot see.

Productive Interference

Page’s deeper insight is that diverse models don’t just add—they interfere. In physics, interference is what happens when waves overlap: sometimes they amplify, sometimes they cancel. The same thing happens with mental models.

Consider the quant funds again. A statistical-arbitrage model sees patterns in price data and bets on their continuation or reversal. An ecology model asks: how many other predators are hunting the same prey? A network model asks: what happens when connected nodes fail simultaneously? A reflexivity model asks: does the act of trading change the pattern being traded?

Each model alone gives a partial, potentially misleading signal. The statistical model says the trade is attractive. The ecology model warns that crowded trades have negative expected value precisely because they are crowded. The network model flags that correlated positions create systemic vulnerability. The reflexivity model notes that the very success of a strategy attracts capital that degrades it.

Held together—allowed to interfere—these models produce something richer. The statistical signal doesn’t disappear; it becomes a starting point, to be checked against questions it cannot answer on its own. The interference pattern is the insight.

This is not the same as saying “consider multiple perspectives”—a piece of advice so generic it communicates nothing. It is a structural claim: if you only ever generate one kind of prediction, you forfeit the error-canceling benefit that an ensemble of approaches can provide. In complex systems, that forfeiture is costly. The ceiling is set by the architecture of your approach, not just the effort you put into it.

Why This Matters Now

There is a reason multi-model thinking is becoming more urgent, and it has to do with the tools we are now working alongside.

AI executes within a framework with consistency and speed that humans can’t match. Within its domain of validity, it often outperforms us. But here is what AI doesn’t do: it doesn’t know when its model applies. It generates confident answers whether the framework fits or not.

This means the human value-add has shifted. It is no longer primarily about executing within a framework—the machine does that better. It is about knowing which framework applies to the situation at hand, and recognizing when you have crossed the boundary into a domain where the model breaks.

This is multi-model thinking. And it is now the scarce resource.

Research is beginning to confirm this. Studies of knowledge workers using AI find that performance improves dramatically for tasks inside the AI’s competence—but degrades, sometimes severely, for tasks outside it. What matters is not just domain expertise but workflow judgment: when to lean on the tool and when to override it.

After Garry Kasparov lost to Deep Blue, he popularized “centaur chess”—human-AI teams competing against each other. The striking result, observed in subsequent freestyle tournaments, was that the best centaurs weren’t the strongest chess players. They were people with good judgment about when to follow the machine and when to deviate. Weak humans with strong process beat strong humans with weak process.

If AI commoditizes single-model execution, then the skill that remains scarce is the ability to work across models—to hold frameworks loosely enough that you can switch between them as the situation demands.

What This Series Will Do

Over the coming months, this newsletter will do something unfashionable: borrow. We will take frameworks from information theory and ask what “knowing something” actually means when information is always relative to an observer. We will take ideas from evolutionary biology and ask why strategies decay even when nothing seems to have changed. We will use network science to understand why systems fail in correlated ways, ecology to understand competition for limited resources, and dynamical systems theory to understand why stable systems can suddenly reorganize.

Investing will be our running example, our laboratory, our source of concrete illustrations. But the goal is not to make you a better investor (though that might happen). The goal is to make you a better thinker in any domain where single models reliably fail—which is to say, any domain that matters.

This is harder than mastering one thing. It requires comfort with partial knowledge, tolerance for contradiction, willingness to hold frameworks loosely rather than gripping them as identities. It is less satisfying to the part of us that wants certainty and expertise.

But complex systems don’t care about our satisfaction. They are what they are: phenomena that sit at the intersection of multiple disciplines, resistant to any single framework’s explanatory ambitions. Understanding them requires a toolkit as varied as the phenomena themselves.

The discipline is in asking: what kind of problem could this be?

References & Further Reading

1. The Model Thinker: What You Need to Know to Make Data Work for You: Scott E. Page

2. What Happened to the Quants in August 2007?: Khandani & Lo

3. Navigating the Jagged Technological Frontier: Dell’Acqua, McFowland, Mollick et al.

4. Deep Thinking: Where Machine Intelligence Ends and Human Creativity Begins: Garry Kasparov

5. The Difference: How the Power of Diversity Creates Better Groups, Firms, Schools, and Societies: Scott E. Page