Returns are not the number
How to judge a strategy — including this one — by more than the figure people stick to.
Most people compare strategies by one number: the return. It is the wrong number on its own, and the whole desk is built on that being true. This chapter is the vocabulary for reading the rest of it — and for reading anyone else's track record.
Return is what you got. Risk is what you had to sit through to get it.
Two strategies both make 12% a year. One does it with days of ±0.7%; the other with days of ±2% and a stretch where a third of the money was gone. They are not the same 12%. The second one gets abandoned at the bottom by most of the people who hold it, which is the point: a return you cannot hold is not a return you will collect.
Volatility is the size of the typical day, annualised. The S&P 500 runs at about 19% a year, which in practice means one-day moves of 1% are ordinary and 3% happen every year. Drawdown is the fall from a peak to the next trough; the worst drawdown is the deepest one on record. It is the number that decides whether you keep the strategy, because it is the number you feel.
Sharpe: return per unit of risk
The Sharpe ratio is the annual return divided by the annual volatility. It asks: for every unit of swing you sat through, how much did you get paid? The index sits around 0.5–0.8 over long stretches. A 1.0 means the return matched the risk one for one. Serious systematic desks live between 1 and 2 and treat 3 as a claim to be checked. It is the first number a professional asks for, before the return, because it is the one that says whether the return was skill or exposure.
Sortino is the same ratio counting only the downward swings — the ones you mind. It is higher than Sharpe for anything that goes up in jumps and down in steps, which is what you want. Profit factor is total winnings divided by total losses on closed trades; 1.3 means the wins are a third bigger than the losses in aggregate. Hit rate is how often a trade finishes up, and it is the most overrated of all of these: a 40% hit rate with wins twice the size of losses is a good business, and a 70% hit rate with occasional large losses is how accounts die.
Alpha and beta: which part was yours
Most of any stock's move is the market. Beta is how much of the market a position carries: 1.0 means it moves with the index, 0.35 means a third as much. Alpha is the return left over after the market's share is taken out — the part the strategy actually earned. A book that returns 14% with a beta of 1 in a year the index made 14% earned nothing; a book that returns 12% with a beta of 0.35 in that year earned about 7% on its own. The site's 0–100 Alpha score is this idea applied to one name: the market subtracted, the remainder ranked.
The reason this matters commercially: anyone can have beta for free by buying the index. What is worth paying for is a return that does not depend on the index going up — and the only way to see whether a strategy has one is to look at beta and alpha, not the headline.
Why the smaller swing compounds better
Losses cost more than gains earn. Down 20% then up 20% leaves you at 96, not 100; down 50% needs up 100% to get home. Over years, the strategy with the smaller swings compounds more of its average return into actual money, which is why two strategies with the same arithmetic average can end up far apart. Volatility is a tax on compounding, and drawdown is the bill.
This is also why a lower-volatility book with a good ratio is the better base for seeking return: if you want more, you size up a stable engine — and take the drawdown that comes with the size, knowingly — rather than loosen the rules of an unstable one. A 1.0-Sharpe book at 12% volatility run at 1.6 times size is a 19% book at the index's volatility with a fraction of its worst drawdown. The ratio is the asset; size is the dial.
The worked example: the desk's rules against buy-and-hold
Replayed 2019 to September 2026, after costs, twenty slots, the engine's own rules as they will ship in the next generation. Buy-and-hold is the S&P 500 over the same days.
| the book | buy & hold | |
|---|---|---|
| annual return | 12.3% | 14.6% |
| volatility | 12.1% | 19.4% |
| Sharpe | 1.01 | 0.75 |
| worst drawdown | −11.4% | −41.7% |
| 2022 | −7% | −19% |
| beta to the index | 0.35 | 1.00 |
| on days the index fell more than 1% | −0.86% | −1.99% |
Read by the return alone, buy-and-hold wins. Read by the whole table, the book keeps most of the return at two-thirds of the swing, a quarter of the worst fall, and a third of the market exposure — and its 7% a year of alpha comes from a source only about half correlated with the index. Put half in each and the combination returns more than the book, with a worst drawdown of −25% instead of −42%. That is what a second engine is for.
What this table does not say, and the desk will not claim: that the book beats the index in a bull year. In 2023, 2024 and 2025 the index simply outran it, and there are stretches — a summer rotation, for one — where the book loses while the index rises. A strategy that promises to beat buy-and-hold every year is selling you its beta. This one promises most of the return with a fraction of the pain, every rule in the open, and a live record you can check on the engine page.
How to read anyone's track record
Ask for the volatility and the worst drawdown next to the return. Ask for the beta — if it is near 1, the return was the market's. Ask whether the numbers are after costs and whether the period includes a bad year; a record that starts in 2023 has not met a bear market. Ask whether the rules were fixed before the period they are shown on, or fitted to it. Ask for the live record, separately from the replay, and how long it is. The desk answers all six on the engine page; the sixth is the youngest, and it is the one that will matter most.