Sharpe ratio, drawdown, and profit concentration

Return per unit of volatility — how much of the growth came from a smooth, repeatable process versus a wild ride that happened to end up positive. Two strategies can both show 300% growth: one earning it in small, consistent steps (a healthy Sharpe ratio, comfortably above 1), the other earning most of it in one or two enormous trades against a mostly flat or losing base (a Sharpe ratio near zero). The growth number can't tell them apart. The Sharpe ratio is built specifically to.

A Sharpe near zero on a strategy with strong headline growth is usually the clearest single tell that the number is doing more work than the strategy is.

The largest drop from a peak in account equity before it recovered — how bad the worst stretch actually got, lived through in real time rather than smoothed out by the final growth chart. A strategy showing 40%+ maximum drawdown has already demonstrated it can lose nearly half an account's value; nothing about a subsequent recovery makes that less likely to happen again. Drawdown is one of the few statistics in trading that tends to repeat, because it's usually a property of the strategy's risk-sizing, not a one-off event.

What share of total profit came from the best trade, or the best short window of days. Many platforms flag this directly — "80% of growth achieved within 3 days" is a real, common warning, and it means exactly what it says: take that one window out, and the track record looks very different. A strategy whose entire growth chart rests on a handful of trades isn't demonstrating an edge so much as reporting the result of one bet that worked.

The strongest check of the four, and the hardest for a seller to fake. Every strategy performs well on the data it was built and tuned against — that's not evidence of anything. What matters is how much of that performance survives on data the strategy never saw during development. A strategy that loses most of its profit factor moving from in-sample to out-of-sample was fitted to noise in its own history; one that holds up close to its original numbers has something more likely to be real.