Guides

From Backtest to Trading Plan: Position Sizing and Risk Per Trade

A positive expectancy tells you a setup is worth trading. It doesn't tell you how much to risk on any single trade — and getting that wrong can turn a genuinely good edge into a blown account.

Stratlog Guides · Risk Management

Suppose your backtest is done. You've logged enough occurrences, and the numbers are good — positive expectancy, a win rate you trust, an average win comfortably bigger than your average loss. It's tempting to treat that as the finish line. It isn't. A backtest tells you a setup has an edge. It says nothing about how much of your account to put behind any single occurrence of it — and that decision matters just as much as the edge itself.

Your backtest gives you inputs, not a plan

Win rate, expectancy, average win, and average loss are all inputs to a decision you still have to make. Two traders with the identical backtested setup and identical expectancy can have completely different outcomes trading it live, purely because of how much they risked per trade. Position sizing is the layer that sits on top of a backtest and determines whether a real edge translates into a survivable, growable account — or a fast way to blow one up.

Why a fixed dollar amount doesn't work

Risking a flat $100 on every trade seems simple, but it breaks down as an account changes size. On a $2,000 account, $100 is 5% of capital — aggressive. On a $20,000 account, the same $100 is 0.5% — overly conservative. A fixed dollar amount also doesn't adjust as an account grows or shrinks, so after a losing streak you're risking a larger share of what's left than you intended, right when consistency matters most.

Risk per trade as a percentage of account

The standard fix is sizing every trade as a percentage of current account value, typically somewhere in the 1–2% range per trade for most discretionary strategies. This has a specific purpose: surviving a losing streak without doing lasting damage to the account.

Why 1-2%, specifically: at 2% risk per trade, ten consecutive losses — a bad but plausible run — costs about 18% of the account, assuming no compounding recovery needed. At 10% risk per trade, the same losing streak is functionally account-ending. The percentage isn't arbitrary caution; it's sized around what a real losing streak, the kind your backtest already showed you, actually costs.

Using your own backtest to size with confidence

This is where the numbers from your backtest stop being abstract and start being directly useful. Instead of picking a risk percentage out of habit, you can size it around what you've actually measured:

Find your max consecutive losses

Look back through your logged occurrences and find the longest losing streak that actually happened. This is more useful than a theoretical worst case — it's a real run your setup has already produced.

Multiply it by your intended risk percentage

A losing streak of 8, at 2% risk per trade, costs roughly 16% of the account (before compounding effects). Ask honestly whether that's a drawdown you could sit through without abandoning the strategy.

Adjust risk per trade until the answer is yes

If an 8-loss streak at 2% risk feels unacceptable, lower the risk percentage until the same streak produces a drawdown you could genuinely tolerate without changing your behavior mid-streak.

Recalculate as your sample grows

Your worst losing streak so far is not guaranteed to be your worst ever. As you log more occurrences, revisit this number — a strategy that's only been tested through calm conditions may show a longer streak once more data comes in.

What position sizing can't fix

It's worth being direct about the limits here: position sizing manages how painful a losing streak is. It cannot turn a negative-expectancy strategy into a profitable one. If your backtest shows negative expectancy, no risk-per-trade percentage makes that setup worth trading — the fix is a different setup or different rules, not smaller position sizes on a broken edge.

✕Increasing size after a losing streak to "win it back." This is the opposite of what your backtested max-consecutive-losses number is for — it exists so you can survive a streak at your planned size, not so you abandon the plan when the streak actually happens.
✕Sizing based on conviction instead of the plan. A setup "feeling" stronger than usual isn't information your backtest captured — if the rules are the same as every other logged occurrence, the size should be too.
✕Never revisiting risk percentage as the account changes. A risk percentage chosen for a $2,000 account may need reassessing at $20,000 — not because the math changes, but because the psychological weight of a drawdown often doesn't scale linearly with the trader managing it.

From plan to live execution

With expectancy validated and a risk-per-trade percentage chosen deliberately, what's left is executing the plan with the same discipline that made the backtest valid in the first place — sizing every occurrence the same way, not adjusting based on how a trade feels in the moment. That consistency is the same discipline this whole process has depended on from the first logged row.

STRATLOG tracks the exact numbers this plan depends on — win rate, expectancy, and the PnL amounts behind average and max win/loss — calculated automatically as you log real occurrences.