Position Sizing: The Variable Most Traders Underestimate

If you ask most traders what their edge is, they will describe an entry signal — a pattern, a crossover, a catalyst. What they rarely describe is how much they bet, even though that decision drives more of their long-run outcome than almost anything else.

The mathematics of ruin

Imagine a strategy with a 55% win rate and average win equal to average loss. That is a real edge. Now run two versions:

• Version A: fixed 2% risk per trade
• Version B: variable sizing, sometimes 10%, sometimes 1%, based on “feel”

Both have the same signal quality. Version A compounds steadily. Version B will, over enough trades, hit a string of large positions during a losing run — and the resulting drawdown may be unrecoverable, not because the signal failed, but because the sizing was wrong. This is the mathematics of ruin. It does not require a broken strategy; it just requires inconsistent sizing.

The Kelly Criterion and its limits

The Kelly Criterion sizes positions to maximise the geometric growth rate — the theoretical optimum. In practice, “full Kelly” produces terrifying drawdowns because it assumes your edge estimate is perfectly calibrated. It almost never is. Most serious systematic traders use a fraction of Kelly (typically quarter- or half-Kelly) to reduce drawdown while preserving most of the growth benefit.

Fixed fractional sizing

The simplest robust approach is fixed fractional sizing — risk a constant percentage of account equity per trade, regardless of conviction. At 1% per trade, you need 70 consecutive losses to halve your account. That is almost never what ends a strategy. Oversizing on “high conviction” trades is.

What good tooling does

A backtesting engine that runs the same strategy across different sizing rules will often reveal that the signal is fine but the historical sizing assumption was distorted. Pulsar makes sizing an explicit, front-of-screen parameter — not something left implicit in the strategy logic.

This content is educational only — not investment advice.