Risk Update: October 2026 – “The Geometry of Winning”

“Uncertainty must be taken in a sense radically distinct from the familiar notion of risk, from which it has never been properly separated. The essential fact is that ‘risk’ means in some cases a quantity susceptible of measurement, while at other times it is something distinctly not of this character; and there are far-reaching and crucial differences in the bearings of the phenomena depending on which of the two is really present and operating. It will appear that a measurable uncertainty, or ‘risk’ proper, as we shall use the term, is so different from an unmeasurable one that it is not in effect an uncertainty at all.” Frank Knight, 1921, “Risk, Uncertainty, and Profit”.

Risk, Uncertainty, and Profit

“Uncertainty, unknowledge, is what confronts the chooser of action when his act of choice is going to be an experiment the making of which will destroy the possibility of ever making that experiment again. In such a case we cannot say what will happen, even if we only claim to it half-heartedly, as a ‘probability’. We can only attain some notion of the kinds of thing that can happen.” GLS Shackle, 1949, “Expectation in Economics”.

Expectation In Economics

These are not new issues for us. We used both these quotes back in one of our all-time favourite Updates, the March 2023 Update – “Probability vs Possibility” Convex Strategies | Risk Update: March 2023 – Probability vs Possibility. In one way or another, all of our writing is some version of this issue. This is the core of what we designate as Sharpe World; the consensus world of finance and economics where simplistic tools of probability have been adopted and then applied to a world where uncertainty reigns. We have a whole string of catchphrases around this topic.

[For those interested in the latest recap of the failures that led to the Silicon Valley Bank collapse, which we discussed at the time in our “Probability vs Possibility note, we refer you to this recent review from Tom Hoenig SVB: One Bank’s Failure Became a Systemic Crisis. It is a good review which gets into the endogenous/moral hazard/regulatory nature of the collapse]

“Risk isn’t what you think is going to happen, it is what hurts IF it happens.”

“History alone is a poor measure of risk.”

“Banks don’t go out of business taking risk; they go out of business leveraging things that they can account for as riskless.”

“The only true alpha is convexity. Anything else is just too short a time series.”

The true dangers, the dark corners of systemic risk, are not the risks that Sharpe World measures; it is the risk that they are permitted to not measure.

Our friend Jeremy McKeown wrote a wonderful note on precisely these topics, at his Hypernormal Times Substack, titled “Manufactured Certainty”. Jeremy digs into the acceptance in the world of physics that uncertainty is ubiquitous and is something to be accepted, not something to be overcome.

Manufactured Certainty – by Jeremy McKeown

“Having surrendered certainty, not as a temporary concession pending better equipment, but permanently and as a matter of principle, physics proceeded to become the most predictively successful discipline in human history… The lesson is, once you stop insisting on definite values that aren’t there, you can establish what is if you follow the right methodology.” Jeremy McKeown, September 2026.

The high priests of modern economics and finance, aka Sharpe World, went the other way. They chose simplified mathematical assumptions of linearity, Gaussian Normality, time-ignorant ergodic slices of time, all leading to precision and a representation of knowability, of certainty. Unfortunately, the result is that they are precisely wrong. That wrongness is then hardcoded into functions that come without skin-in-the-game and the necessary role of tacit knowledge needed to manage uncertainty. What we end up with is a self-organizing criticality that creates those very tails that we are told don’t exist.

Per Jeremy’s note, to compete, Sharpe World entities must accrue the benefit of being exposed to the tails and, subsequently, necessitate the bailing out of the regulated system. Wash, rinse, repeat.

“The one reform requiring banks to characterise their tail risks has been postponed for the better part of two decades after the crisis that demonstrated why it is necessary, because an institution obliged to describe its tail honestly is at a disadvantage against one that is not. Manufactured certainty is no longer an intellectual error. It is a measurable competitive advantage, and the authorities have now publicly priced it as one.” Jeremy McKeown, September 2026.

Jeremy sums up, very succinctly, what we would class as a very good investment/risk management mindset.

“Judge a position by what happens when you are wrong, not by how likely you are to be right… Prefer a known premium paid for an unknown parameter over unbounded exposure to it.” Jeremy McKeown.

When we get into a world where measured risk is all that matters, we inevitably find ourselves exposed to the perils of back tests, black boxes, replicators, volatility laundering, etc. The entire premise of the Sharpe World risk construction is based on the concept of extrapolating (recent) historical data series, a measure of frequency, into probability distributions and then treating the risk as though it was no different than the known probabilities of a casino game of chance. One need not wait long in a meeting with Sharpe World practitioners before you will hear probabilities of future occurrences being quoted with absolute precision.

For years now we have been using our image of the Football/Soccer Pitch, overlaid with a Gaussian/Normal distribution and a Shannon’s Entropy Curve, to help visualize the respective impacts of frequency versus magnitude.

Figure 1: Frequency vs Magnitude. Contribution to S&P 500 Long Term CAGR over 40yrs

Source: Bloomberg, Convex Strategies | Risk Update: June 2023 – One Thing

The black Normal Distribution represents the frequency of where the ball spends time, or the frequency of monthly returns in the 480-month data series of the S&P 500. On average, 50% of the time, the ball is in the middle of the pitch, and, within a couple of standard deviations, the ball spends its time in between the respective penalty boxes. The Shannon’s Entropy Curve measures the impact of where the ball spends time on the outcome of the match. The shading represents the contributions to the long-term CAGR (Compounded Annual Growth Rate) by percentiles. The 10 worst months (2 percentile worst performing months) are shaded in orange and contribute circa 40% of the long-term CAGR. The 10 best months (2 percentile best performing months), shaded in blue, contribute circa 30% of the long-term CAGR. The middle 460 months (96 percentile of months) contribute less than 30% of the long-term CAGR.

As Claude Shannon told us, there is no information in the expected outcome. Football strategies, like investment strategies, should be built around resilience away from the mean, in both directions, not optimized to the mean. Bound, to the best extent possible, the downside so that you can aggressively pursue the upside. Jumping analogies, put good brakes on your car so you can, safely, drive faster. We wrote about resilience in our so-named August 2025 Update Convex Strategies | Risk Update: August 2025 – “Resilience” and linked to this note from Chen-Embrechts-Wang on super-Pareto distributions.

2025Chen-Embrechts-Wang-IME.pdf

It is a wonderful note on the very known dynamic of power-law distributions, heavy tailed distributions, that make it quite clear why one should avoid optimizing to the simplified assumptions of Sharpe World. In such fat tailed type distributions there, literally, is no definable mean! Making assumptions about probability distributions and mean-variance optimization is, pun not intended, meaningless.

“…for super-Pareto losses, the action of diversification increases the risk uniformly for all risk preferences, such as VaR, expected utilities, and distortion risk measures, as long as the risk preferences are monotone and well-defined. The increase of the portfolio risk is strict, and it provides an important implication in decision making: For an agent who faces super-Pareto losses and aims to minimize their risk by choosing a position across these losses, the optimal decision is to take only one of the super-Pareto losses (i.e., no diversification).” Chen-Embrecht-Wang, June 2025.

This gets right at the heart of why we say that it is the assumed diversification benefit, the assumed stable correlation, that is generally the greatest unmeasured risk in most multi-asset portfolios. To the extent that the correlation assumption allows some amount of explicit or implicit leverage, i.e. additional risk to sit on top of loss absorbing capital beyond what otherwise would be the case, this then evolves into the self-organizing state of market fragility.

Physicist, cum quant-finance legend, JP Bouchaud wrote this recent fantastic note on just this topic, “Non-Equilibrium Economics: A Physicist’s Point of View”. JP discusses the flawed perspective of the Sharpe World crowd of stable equilibriums as opposed to the realistic view of self-organizing endogenous systems forever chasing an unattainable strange attractor (to borrow Lorenz’s terminology). JP’s list of 118 references for this short paper is enough on its own to make this note worth printing out and keeping close to hand. A small sampling of some of our favourites that appear include Per Bak, Doyne Farmer, Brian Arthur, Nassim Taleb, Hyman Minsky, Herbert Simon, Andrew Haldane, Xavier Gabaix – all frequent contributors to our own works.

Non-Equilibrium Economics: A Physicist’s Point of View

“In this review I argue, from a physicist’s vantage point, that the concept of equilibrium – so central to mainstream economic thinking – is likely to be the exception rather than the rule in large, complex, interacting systems… I show how excess volatility, endogenous crises and crashes, inflation swells and persistent inequalities can all emerge naturally from genuinely out-of-equilibrium dynamics, without invoking large exogenous shocks. Three generic mechanisms recur throughout: trapping in a multiplicity of history-dependent equilibria; the impossibility of dynamically reaching equilibrium, leading to oscillations and chaos; and the spontaneous evolution towards fragile, marginally stable states – the self-organized criticality paradigm.” JP Bouchaud, July 2026.

JP touches on so many of the things that we hold important in the disconnect between the Sharpe World equilibrium view of things and the reality of complex, interactive, systems. Time and history matter. How participants/agents learn has implications to system stability. Optimization leads to fragility. Efficiency and resilience are mutually incompatible.

“… the quest for efficiency and the necessity of resilience may be mutually incompatible. Optimization, in a complex system, tends to consume the very redundancy that would protect it. Fragility and efficiency are two sides of the same coin.” JP Bouchaud, July 2026.

This goes back to our discussions in the July 2026 Update – “Rogers’ Paradox” Convex Strategies | Risk Update: July 2026 – “Rogers’ Paradox” and the necessary balance between individual learning, through trial and error (tȃtonnement, as JP refers to it), and social learning. Too long in a steady state risks too much acceptance of established social learning, leading to fragility against dynamic adjustments. Another of our inhouse axioms:

“Optimization is the art of making a system perfectly adapted to the last regime—and therefore maximally vulnerable to the next.”

From a central banker’s perspective, one might argue that they optimized their models and actions to the past regime of abnormally low volatility of their price stability measures and unprecedentedly low interest rates. It is easy to imagine that they now might find themselves vulnerable to a new regime. There is at least one central banker (sadly, outgoing) that agrees with us – Isabel Schnabel of the ECB. Below is a speech recapping her time at the ECB as she prepares to depart.

Navigating stormy waters: reflections on monetary policy

“We have moved from a world in which inflation was persistently too low to one marked by recurrent inflationary shocks; from a historically accommodative monetary policy stance to the fastest tightening cycle in the history of the ECB; and from a deeply integrated global economy to one increasingly shaped by fragmentation and geopolitical tensions.” Isabel Schnabel, November 2026.

It is a good speech which shows that she has indeed learned a few things. She lays it out as four takeaways.

  1. “Preserve room for manoeuvre by avoiding commitments.”
    1. Recognize regime shifts
    2. Savings glut has switched to a bond glut, aka Hunger Games.
    3. Unpredictable, expect the unexpected.
    4. Forward guidance is a bad idea.
    5. QE has long term consequences and should be used as a stability tool, not a monetary tool.
  2. “What matters is how shocks affect the inflation outlook.”
    1. Their job is to manage inflation, not to explain what types of inflation matter.
  3. “Protect monetary dominance.”
    1. Avoid financial dominance.
    2. Avoid fiscal dominance.
    3. Be more innovative and productive.
    4. Keep r < g.
  4. “Implement monetary policy efficiently.”
    1. The balance sheet is not an efficient monetary policy tool.
    2. Banks should assess their own liquidity needs.

All seems pretty obvious but such recognition is exceptionally rare amongst her colleagues. Given their lack of response to the current price stability challenges in their most recent policy meeting, we will withhold our optimism that actual change is taking place at the ECB.

Figure 2:Eurozone HICP Index (white). 1997 – Oct 2026. Circa 1.6% per annum trend 1997 – July 2021 (red). Circa 4.4% per annum trend July 2021 – Oct 2026 (green)

Source: Bloomberg, Convex Strategies

Of course, the ECB is not the only institution that seems to have an asymmetric bias as to which side of their price stability mandate they are willing to tolerate a certain amount of divergence. The Fed, likewise, chose to hold their policy rate unchanged at the October policy meeting.

Figure 3: US CPI Index (white). 1997 – October 2026. Pre-FAIT Trend to August 2020 2% (red dash). Post-FAIT Trend from August 2020 4.36% (green dash)

Source: Bloomberg, Convex Strategies

Kevin Warsh has regularly said that inflation is a choice. Makes one wonder what it is that he and his colleagues are choosing?

As ever, we would advocate resiliency. Stop managing to forecasts and expectations. Start bounding the downside, the bad outcomes, allowing for more opportunity to pursue the upside, the desirable outcomes. Our construction of the Convex Goalkeeper Theorem for the World Cup this year gives great real-world examples of the benefit of resilience.

Research | Research

The single-elimination stage of the World Cup provides an excellent environment to apply these concepts because, like the reality of the worlds of finance and economics, it represents a non-ergodic environment where sequencing and actual path matters. The objective is winning the tournament, if you lose a game, you are out. You can’t recover. Ensemble averages are meaningless. It is very like the compounding path of an investment strategy.

We recently gave a presentation comparing the World Cup to various investment strategies. We titled the presentation “The Geometry of Winning”. The premise of our model is that it measures Defensive Magnitude (DefMag) and Offensive Magnitude (OffMag) and calculates Expected Goals (xG) for both teams. What it is trying to capture is exactly what we show in our above Football Pitch visualization, that it is resilience away from the mean, that it is magnitude in the respective penalty boxes, that will drive the outcomes of the tournament. Off this we calculate a ConvexScore. For what it is worth, Spain (the eventual winner) had the dominant ConvexScore throughout the tournament. 

We thought we would share a few of the examples here.

First up, the epic semi-final match between England and Argentina. Infamously, England took the lead in the 55th minute and then decided to ‘park-the-bus’. After England gave up on all attacking play for the final 1/3rd of the match, Argentina’s superstar, Lionel Messi, did what great players do – he made a string of accurate crosses until, eventually, Argentina won the match with two goals late, in the 84th and 90th + 2 minute points.

Figure 4: Convex Goalkeeper Theorem: England 1 – Argentina 2 (Semi-Final) aka Foregone Upside

Source: FIFA Technical Study Group Post-Match Summary Reports, OPTA, https://convex-strategies.com/research/

England lost sight of the purpose of a strong defence. You defend to attack. They gave up on their OffMag, while still being exposed to the downside, in this case a particularly dangerous, fat tailed, downside in the form of Lionel Messi.

We compare the ‘park-the-bus’ strategy to the ever present, Sharpe Ratio optimising, prolific strategies in the investment universe. Strategies that optimise to the measured risk of volatility and forget that the objective is compounded wealth. A great example of such a strategy, one that explicitly foregoes the upside to only probabilistically reduce downside, is the good old, covered call strategy.

Figure 5: Buy Write (red) vs Put Protect (blue). Foregoing Upside ≠ Risk Management

Source: Bloomberg, Convex Strategies

We have compared here the CBOE BuyWrite Index (red) to the CBOE PutProtect 3m Index. The BuyWrite owns the S&P and writes and rolls call options, collecting premium in exchange for foregoing the upside. It is the prototypical strategy of managing risk by taxing the up tail. The PutProtect owns the S&P and pays premium for downside put protection. It has explicit risk mitigation and continues to have full exposure to unbounded potential upside. The blue strategy would have a good ConvexScore. Not so much the red strategy.

England’s coach, Thomas Tuchel, who presumably made the strategic call to ‘park-the-bus’, doesn’t see it the same way. He thinks possessing the ball in the mean zone is what it takes to win.

Thomas Tuchel blames ‘English DNA’ for World Cup failure

In this article, Tuchel tries to make the case that England failed because they couldn’t control possession, claiming that it is not in the English players DNA (Tuchel is German) to control the midfield. He even blames them for wanting to attack. Our math supports just the opposite; possession does not win football matches. Defensive resilience that strengthens as the ball approaches your defended goal, paired with offensive aggressiveness that strengthens as you approach your attacking goal, are the paths to victory.

Another very interesting, and rather unique game, was the match for 3rd Place between England and France. This game is unique because it was the only game that could be thought of as ergodic, as in the outcome of this match had no bearing on the path-through-time objective of winning the tournament. Both teams had already been eliminated, losing this game made no difference in the objective of winning the World Cup.

France, in particular, responded to the shift from a non-ergodic match to an ergodic match with a very different style of play. They threw caution to the wind and went all out. Scoring 4 goals, including a 3-goal hat-trick from their superstar, Mbappe. It was a thrilling back and forth game, which France lost to England 6-4.

Figure 6: Convex Goalkeeper Theorem: 3rd Place Game aka The Ergodic Game

Source: FIFA Technical Study Group Post-Match Summary Reports; https://convex-strategies.com/research/

The above chart compares some of our statistics for France’s loss to England with France’s earlier hard-fought victory over Morocco, back when you had to win to move on. A few things standout; 1) despite spending a lot more time in the attacking penalty box, France’s xG was actually lower in the England match (Mbappe, like Messi, is a fat tail), and 2) against Morocco, France gave up just 1 shot on goal, which was saved, while against England they gave up 11 shots on goal, conceding 6 goals. They had disastrous DefMag.

We see this match as a perfect example of the far too common flaw of targeting arithmetic returns and focusing on a single calendar period as an objective. This, in the world of investing, leads to what is commonly known as ‘Volatility Drag’. It is simply the decaying of realized geometric returns below the hoped for expected arithmetic return (aka Expected Return).

Figure 7: Ergodic Assumption = Expected Utility > Actual Utility aka Volatility Drag

Source: Expected-utility maximizers don’t maximize utility. Convex Strategies | Risk Update: May2025 – “Just Do It”

The above visualization comes from the guru of Ergodicity Economics, Ole Peters. It is simply the simulation of a coin-toss game where the player gets a 50% return on his capital if he wins the toss and losses 40% of his capital if he losses the toss. The ‘Expected Utility’/arithmetic return is +5% per toss, represented by the dark blue line over many number of games. The ‘Actual Utility’/geometric return is, approximately, -5% over time. The median player tracks the red line.

Volatility drag is the bane of all investment strategies optimized to expected returns. It is precisely what adding positive convexity to a portfolio is trying to address. It is why folks will regularly hear us declare, when talking about the improved compounding paths of convex-style portfolio, that “it’s just math”.

Finally, to Spain. Spain dominated the ConvexScore throughout the tournament, mainly because they dominated the DefMag but also because they used their DefMag dominance to attack aggressively.

Figure 8: Convex Goalkeeper Theorem: The Champions

Source: FIFA Technical Study Group Post-Match Summary Reports, https://convex-strategies.com/research/

Spain’s stats were extraordinary from their first match straight through the final match against Argentina. In the final, Argentina took zero shots on goal over 120 minutes (only one shot, that was blocked, overall). Spain, however, managed 20 shots, 12 on target. They used their remarkable defensive resilience to attack the opponent’s box. Our model is not a prediction model. It is a model that tells you how games are won. The model did, however, show from the very first game that Spain understood how to win games and how to manage their way to the ultimate objective of surviving and winning the non-ergodic, elimination, tournament.

We would compare Spain’s path to victory with our good old standard convexity transformed Always Good Weather Portfolio (AGW) where we take a simple 60/40 Equity/Bond balanced portfolio, between S&P (SPXT Index) and US Treasuries, and simply adapt it to our Football Pitch visual. Get rid of the 40% in bonds and split it equally between an aggressive goalkeeper (our bootstrapped EurekaHedge Long Vol + WITH Long Vol Indices) and attacking goal scorers (Nasdaq XNDX Index). Below we compare that to the S&P Risk Parity 10%Vol Index, a strategy that optimizes to the mean zone, leveraging the supposed benefits on low vol/low correlation (what we would call a recipe for Volatility Drag).

Figure 9: AGW (blue) vs Risk Parity (red), aka DefMag Optimized vs Mean Zone Optimized. Compounding View.

Source: Bloomberg, Convex Strategies | Risk Update: December 2023 – “It’s Just Math”

In the non-ergodic path to superior compounded returns, it is no surprise that the strategy with superior DefMag and OffMag wins out. And, just like our Football Pitch visual, the key to that winning takes place in the extremist percentile of occurrences. The 1 percentile worst month and the 1 percentile best month drive the bulk of the compounding separation.

Figure 10: AGW (blue) vs Risk Parity (red), aka DefMag Optimized vs Mean Zone Optimized.

Source: Bloomberg, Convex Strategies | Risk Update: December 2023 – “It’s Just Math”

It is the resilience in the most treacherous period that allows for the aggressiveness in the period of greatest upside opportunity. It is the confidence in the brakes that allows racing teams to explore ways to go faster. The link below is to a session at the recent Simplify Entering the Fall Conference in NY. The star of this clip is one Max Taylor. Max is a junior circuit Indy race car driver, just on the cusp of the possibility of getting promoted to the full Indy professional level. We were lucky enough to get an invite to participate in this conference, as well. Despite what you hear from young Max in this interview, we honestly did not rehearse or script him in his comments!

Entering the Fall 2026 | Risk Management at 200 MPH

Max does a sensational job in this interview (his father, Mike, is pretty good too!). We would highlight two key points that Max makes:

  1. You must prepare the car BEFORE the race.
  2. His objective is to win the season long, multi-race, championship. He is all about compounding.

One last thing to highlight, again going back to our “Resilience” Update. We put this quote in that note from the exceptional book from Judith Rodin, “The Resilience Dividend”. It makes the point that we continuously hammer on about. The resiliency of a strong goalkeeper allows you to attack. Better brakes allows you to explore ways to go faster. Better explicit risk mitigation/convexity allows you to open more right tail opportunities. It is not simply that the goalkeeper/brakes/convexity could save you from doom (though, that alone, is reason enough for utilizing them), it is the things they allow you to do, it is about the “resilience dividend”.

“What is resilience? Resilience is the capacity of any entity – an individual, a community, an organization, or a natural system – to prepare for disruptions, to recover from shocks and stresses, and to adapt and grow from a disruptive experience. As you build resilience, therefore, you become more able to prevent or mitigate stresses and shocks you can identify and better able to respond to those you can’t predict or avoid. You also develop greater capacity to bounce back from a crisis, learn from it, and achieve revitalization. Ideally, as you become more adept at managing disruption and skilled at resilience building, you are able to create and take advantage of new opportunities in good times and bad. That is the resilience dividend. It means more than effectively returning to normal functioning after a disruption, although that is critical. It is about achieving significant transformation that yields benefits even when disruptions are not occurring.” Judith Rodin, “The Resilience Dividend”, 2014.

Read our Disclaimer by clicking here