The Asymmetry Edge
A $40 million Apache taken down by a $20,000 drone is asymmetry. So is a $1,000 put option that pays $40,000 in a crash. Ed Thorp built a career on this idea. The market is currently sleepwalking past it.
Apache vs drone
caps $30k downside
investing method
tolerance for ruin
David and Goliath, oil, and a market that isn't listening
Drones are rewriting warfare. A $40 million Apache helicopter taken down by a $20,000 drone is not a one-off, it is the new equation. A 2,000x cost asymmetry, with the cheap side winning. The implications run through defence budgets, insurance, geopolitics, and supply chains. This is the kind of structural shift markets typically take years to price.
Oil keeps flowing, with China and the US keeping the wheels turning through a combination of strategic reserves, OPEC manoeuvring, and quiet diplomacy. The question is how long the choreography can continue before something gives.
The market itself, meanwhile, is acting as though none of this matters. Indices keep grinding higher. Conviction is high. Hedging is cheap. That combination has shown up before, and rarely ended well.
The interesting question is not whether the market is wrong. It is whether the cost of being prepared for the case where it is wrong has ever been lower than it is now.
Asymmetry: a lot of upside, very little downside
Asymmetry means a lack of equality between two sides of a relationship. In markets, what you want as an investor is a payoff structure where the upside is large and the downside is small. Not a 50/50 bet on direction. A skewed bet where being wrong costs you a little and being right pays you a lot.
The cleanest way to see this is with a worked example. Imagine you have $100,000 invested long. You think the market will rise around 10%, but you are aware that if you are wrong it could fall 30%. The job of asymmetry is to make sure you survive both outcomes, and ideally profit from each.
A long position plus a $1,000 put hedge
Long $100k portfolio. Buy a protective put for $1,000 that pays out if the market falls past a strike price. Two scenarios:
The hedged line stays above zero in both the +10% and -30% scenarios. The unhedged line is exposed to the full crash.
The structure is the entire point. You spend $1,000 to convert a binary directional bet into a profitable trade in both outcomes. You do not need to be right about direction. You need to be right about the existence of the tail risk and willing to pay a small fee to neutralise it.
Now look at the wave of AI IPOs hitting the market. The implicit bet investors are taking is the opposite of asymmetric. The upside is real but capped by valuation. The downside is real and unbounded if the new era narrative breaks. There is no $1,000 hedge embedded in a Nvidia long. There is just a long.
Ed Thorp: a man for all markets
If you want a framework for thinking about edge, asymmetry, and ruin, Ed Thorp is the source. He counted cards at blackjack, then turned the same approach on the options market and ran one of the highest-Sharpe hedge funds in history. His book A Man For All Markets is the cleanest distillation of how to think about this.
For me, blackjack was a game of math, not luck.Ed Thorp, A Man For All Markets, p.93
Thorp's worldview was that risk, reward and probability are the same problem dressed in different clothes, and most people do not understand any of them well enough. That is true at the blackjack table, and it is true in the equity market. The solution is method.
- Identify a clear edgeIt must be obvious and uncomplicated. If you have to talk yourself into it, it is not an edge.
- Size positions with KellyAllocate based on the Kelly Criterion. Match bet size to the strength of the edge, not your conviction or excitement.
- Keep it simpleSmallest number of side effects, minimum hidden complications. Complexity hides risk.
- Avoid ruin at all costsSurvival is non-negotiable. A blown account cannot compound back from zero.
- Calculate expected profit versus worst caseDynamically, one bet at a time. The right size today is not the right size tomorrow.
Step 4 is the one most investors get wrong. Thorp's whole career rested on never allowing a single position, no matter how attractive, to threaten the existence of the portfolio. Asymmetry without ruin avoidance is just a fancy way to blow up slowly.
What an actual edge looks like in data
Thorp's central insight on markets: as prices change, the odds change. What was a good bet at one price is a poor bet at a higher one. Most investors ignore this. They buy what has been working and assume what is working will keep working. That is momentum thinking, and it eventually meets mean reversion.
The Shiller CAPE ratio is one of the cleanest illustrations of price-as-edge in the data. It does not tell you what the market will do in the next year. It tells you, with remarkable consistency over a century of data, what to expect over the next decade.
Each dot is a starting CAPE level vs the real annual return delivered over the subsequent decade. The relationship is noisy at any one point, robust across the sample.
At a CAPE of 10, the next decade has historically returned around 10% real per year. At a CAPE of 30, that drops to around 3%. At today's CAPE of 38, the line of best fit is sitting near 1% real per year. That is the prior. Not the prediction, the prior.
This is the part that requires discipline. CAPE does not say sell. It says: if you are buying broad US equities at this price, the next decade is unlikely to look like the last one. That is not bearishness, it is arithmetic. The strategy that fits a 1% expected return environment is different from the one that fits 10%.
It also means timeframe matters more than almost anything else. A CAPE-based edge is useless over twelve months and meaningful over ten years. Pick the strategy that matches the horizon, not the other way around.
The bell curve lies. The tails are where you get killed.
In 1900, Louis Bachelier published the first mathematical model of stock prices. He assumed price changes followed a Gaussian distribution, the familiar bell curve. The maths is clean, the implications are tidy, and the model is wrong. Thorp picked this up early. Most of finance still has not.
Log scale on the y-axis. The bell curve says crashes of 4+ standard deviations are essentially impossible. Reality keeps delivering them.
Under a bell curve, the 1987 crash was a once-in-the-history-of-the-universe event. So was 2008. So was March 2020. The fact that we have had four such events in living memory is the proof that the model is wrong. Real markets have fat tails, and the fat tails are precisely where investors get wiped out.
This is why ruin avoidance is non-negotiable. Position sizing built on a normal-distribution view of risk will eventually meet a real-distribution outcome and not survive the meeting. Thorp sized his bets for the actual world, not the assumed one.
What we like and don't like this week
Symmetric exposure at the top
Unhedged long positions in expensive markets. Buying the AI IPO wave at a CAPE of 38 with no downside protection. The payoff structure is bad even before you ask whether the underlying business model holds.
Hedging in all its forms
Late-cycle conditions are exactly when Black Swans show up. Hedging is cheap when nobody wants it, expensive when everyone does. Building protection now, while volatility is still suppressed, is the asymmetric trade itself.
Building an Asymmetric Portfolio
A full walk-through of how to construct asymmetric exposures across the book. Position sizing with Kelly, structuring put protection, sizing tail hedges, and how TMM thinks about ruin avoidance in late-cycle conditions. Worked examples on real instruments.
Reserve Your SeatYou cannot predict the crash. You can size for it.
Thorp's edge was never that he knew what was coming. It was that he had calculated, in advance, what he could afford to lose if he was wrong, and sized every position accordingly. That made him robust in any single outcome and powerful in expectation across many.
Most investors do the opposite. They size up when conviction is high, which is precisely when the market is most likely to surprise them. The discipline is to size positions based on what could go wrong, not what you hope will go right. Kelly, not enthusiasm. Math, not luck.
The Three Wells exist for the same reason Thorp's method works. Diversified across strategies, sized to survive any single failure, and built around the principle that no one bet should be able to ruin the structure. That is what asymmetry looks like at the portfolio level.
