Size Discipline: The Spread That Says Rule or Emotion Is in Charge

Two traders with identical setups and identical win rates can produce wildly different equity curves based purely on how they size, and the sizing decision is the one the trader fully controls. The math is uncompromising: the spread in a trader’s per-trade risk is one of the cleanest predictors of long-term drawdown and one of the most reliable signals that risk is being set by emotion rather than by rule. Size discipline is how Gecko measures it.

Why size variance is a behavioral signal, not a strategy choice

A trader on a rule-based position-sizing scheme (a fixed-fractional 1 percent of equity, say) will show a tight, narrow distribution of per-trade risk in dollars. The risk drifts gently as the account compounds; it does not jump on consecutive trades. A trader sizing by feeling will show a wide, lumpy distribution: small risk on the trades they doubt, large risk on the trades they like, outsized risk on the trades they take in revenge. The distribution shape itself is the diagnosis.

The reason this matters is that the strategy never asked for the variance. The variance is the trader’s emotion overriding the position-sizing rule. Every oversized trade increases the contribution of a single outcome to total P&L; every undersized one mutes a real signal. Both push the account further from the expectancy the strategy was supposed to deliver.

The fingerprint in your trade data

The math: a small sizing leak can erase a good strategy

Suppose a strategy has true expectancy of +0.5R per trade and the trader runs a disciplined 1 percent fixed-fraction sizing scheme. Over 100 trades the expected return is roughly 50 percent (in R, scaled to risk-per-trade), with a tight, modelable drawdown distribution.

Now introduce one behavioral leak: the trader sizes 3 percent on revenge trades, which make up 5 percent of total volume. The math: 95 trades at 1 percent and 5 trades at 3 percent, with the 5 revenge trades running a far worse win rate. The expected return drops; the variance of the equity curve roughly doubles; the modeled risk of ruin at any given drawdown threshold rises sharply. The trader did not change strategy or markets. They changed sizing on 5 percent of trades. The account no longer looks like the same account.

How Gecko measures it

The size-discipline axis computes the trader’s per-trade dollar risk (entry minus stop times size), then reports:

1. The distribution’s coefficient of variation (CV = standard deviation ÷ median). Lower is more disciplined.
2. The size ratio of the largest 10 percent of trades to the median, flagged when above 2.5×.
3. The correlation between trade size and prior-trade outcome, flagged when significantly different from zero.

The axis is most useful as a coaching signal: it does not prescribe a sizing scheme, it grades whatever the trader is doing against its own internal consistency.

A worked example

A futures trader uploads 180 closed trades. Median per-trade risk is $250. CV is 0.78. The largest 10 percent of trades risk $850 on average (3.4× median). Correlation between size and prior-trade loss is +0.34 (positive, statistically significant). The diagnosis writes itself: the strategy is being layered with an unwritten rule that says “size up after a loss to recover.” That rule, never written down, is what the equity curve is actually executing — and it is the rule responsible for the trader’s deepest drawdowns of the year.

The fix: write the sizing rule down, and run it

Size discipline does not respond to good intentions. It responds to a written sizing rule applied with the same mechanical reliability as a stop-loss order:

• Pick one sizing scheme and stick to it. Fixed-fractional (1 percent of equity) is the most defensible default; volatility-adjusted variants build on the same idea.
• Compute the trade size in writing before each entry, ideally automated by a calculator that takes equity, stop distance, and risk-percent as inputs.
• Refuse any size override that is not itself a documented sizing rule. The lifetime cost of a disciplined size that turns out to be a winner is much smaller than the lifetime cost of an undisciplined size that turns out to be a loser.
• Audit the per-trade risk distribution monthly. CV that drifts higher is an early warning sign that the rule is being eroded.

What to read next

The cleanest piece on the dollar weight of sizing decisions in a career is the Paul Tudor Jones profile — defense first, size by rule, never average a loser. The Peter Brandt profile is the half-century version of the same discipline. The glossary entry is the one-paragraph reference.

Grade your sizing →Free to start. First 100 trades on the house. No broker connection required.