The Sharpe ratio is a reasonable first-pass scale for return per unit of volatility. It is a bad final judge for algo strategies because it assumes a world that trading rarely provides.

Limitation 1: returns are not normal

Sharpe uses mean and standard deviation. Real trading returns skew negative, have fat tails, and cluster volatility.

Fix: pair Sharpe with tail metrics and drawdown path analysis, not with more indicators.

Limitation 2: short samples create fantasy rankings

With few trades, Sharpe swings wildly. A lucky month can "prove" nonsense.

Fix: enforce minimum sample rules and stability across walk-forward windows.

Limitation 3: Sharpe ignores path and survival

Two strategies can share a Sharpe while one spends months underwater beyond your tolerance.

Fix: add max drawdown, time-to-recovery, and worst-month behavior to the decision table.

Limitation 4: costs and fills can silently inflate Sharpe

If your volatility shrinks because you ignore realistic costs, Sharpe rises for the wrong reason.

Fix: run a stressed cost pass and compare Sharpe deltas, not only headline Sharpe.

What to use alongside Sharpe

Sortino for downside volatility focus (see dedicated article in hub)
Calmar for drawdown burden per unit return
walk-forward efficiency and OOS retention for time discipline

Annualization traps

Sharpe annualization assumes returns scale cleanly with time. That breaks when:

you mix intraday and daily returns incorrectly
your strategy has sparse trades (many zero days)
your variance is dominated by a few outlier days

If annualized Sharpe looks absurd, check the scaling math before you celebrate.

Benchmark mismatch creates fake Sharpe improvement

If you subtract the wrong benchmark return series, you can inflate risk-adjusted measures without improving tradable PnL.

Align benchmark liquidity and leverage assumptions with what you can actually trade.

Sharpe is not a substitute for capacity checks

A strategy can have attractive Sharpe at toy size and become untradable when size interacts with the book.

Pair Sharpe with execution realism and capacity stress (Strategy capacity).