Reading, not predicting

The forecasting graveyard

Every cycle produces the same casualty: the investor who called it. The one who saw the crash coming, positioned for it, was right — and then kept the position through the recovery, waiting for the second act that never came. Being early is the famous failure mode. Staying late is the common one. Both descend from the same error: believing that because you understood one transition, you can predict the next.

Markets are not weather systems that repeat; they are systems that respond to being understood. A pattern known widely enough gets arbitraged into noise or front-run into a different pattern. This is why forecasting track records decay, and why the loudest macro forecasts cluster at exactly the moments they are least reliable. The honest premise, which the best all-weather thinkers accepted decades ago, is that regimes cannot be predicted.

Their response was to diversify across all possible weather simultaneously — permanent allocations to every state of the world. It is a coherent answer, and a humble one. Our view is that it concedes slightly too much. Regimes cannot be predicted. They can, with discipline, be read — and reading is worth more than it sounds, because most of the damage in markets is done not by the storm but by carrying the wrong posture into it after it has already begun.

Orientation, not prophecy

The distinction matters enough to be precise about. A forecast says: volatility will spike in the third quarter. A reading says: the price of insurance is rising, credit has stopped confirming the equity rally, and the trend is bending — the environment has already changed, whether or not the index has noticed. The forecast claims knowledge of the future. The reading claims only attention to the present — a much smaller claim, and a much more defensible one.

Sailors do not predict the sea. They read it, continuously, and adjust the sail rather than the destination. The destination — ownership of exceptional businesses, compounding across decades — does not move. The posture around it should.

Why these four

We distill the environment through four public signals. Not because four is sacred, but because each one is the concentrated opinion of a different market, and it is the disagreements between them that carry information.

The price of fear. Implied volatility is what the options market charges to insure the near future. It is not a forecast either — it is a price, set by people paying real money, and prices are harder to fake than opinions. When insurance is cheap, risk is being carried casually. When it reprices suddenly, someone has stopped being casual.

The price of time. The slope of the yield curve is the bond market’s opinion about the cycle — what the next few years are worth relative to the present. It has preceded every American recession in half a century, not because it is magic, but because it aggregates the time preference of the most unsentimental capital in the world.

The price of risk. Credit spreads are the canary with the best track record. Equity investors are paid to be optimistic; lenders are paid only if nothing goes wrong, which makes them professionally allergic to deterioration. When spreads widen while equities celebrate, believe the lenders.

The confirmation. Trend is the least clever of the four and the most useful: it reports what is actually happening, as opposed to what should be. A market above its long average carries momentum; below it, mean-reversion risk and falling knives. Trend does not explain. It confirms — or refuses to.

Any one of these can be noise. Two agreeing is weather. Three agreeing is a season. And the moments that matter most are the divergences — insurance repricing while the trend still rises, credit deteriorating under a calm surface. The single most valuable thing the panel does is flag the interregnum: the stretch when the old regime has ended and consensus has not yet admitted it.

What a reading changes — and what it must never change

Discipline lives in the boundary. The reading sets the posture of the edge: whether convexity leans defensive, offensive, or simply earns while it waits. It sets the pace at which the buffer deploys. It sets nothing else.

It does not decide what we own. A business either clears the bar or it does not, in every regime. The moment environmental readings begin to justify buying weaker businesses or selling stronger ones, the framework has failed — the weather has been allowed to redraw the map. Themes decide what enters the core. The regime decides only how the edge stands around it.

This division is what makes reading safe to practice. A forecaster who is wrong is wrong everywhere — the whole book was built on the prediction. A reader who is wrong has mispositioned a sleeve, briefly, and the signals that misled will be the same ones that correct. The cost of a bad reading is basis points of posture. The cost of a bad forecast is the portfolio.

We hold no view about next year’s weather. We try very hard not to be wrong about today’s.