Assigning a strategy to every trade turns a journal into a much more useful dataset. But choosing the “best” strategy only by total PnL can simply reward the strategy you traded most often or one that had a single outsized winner.
Build a comparable scorecard
For each strategy show trade count, net PnL, total R, average R, win rate, expectancy, Profit Factor, drawdown and plan adherence. When possible, include medians or distributions to reduce the influence of outliers.
Require a minimum sample
Define a minimum number of trades before using a group for decisions. There is no universal number that guarantees reliability; the goal is to avoid strong conclusions from tiny groups.
Separate system from execution
If a strategy performs poorly but also contains many mistakes, execution may be the issue rather than the setup. Compare trades that followed the plan with trades that did not.
Explore subgroups carefully
You can then cross strategy with timeframe, direction, market, session or event. Every extra filter reduces sample size, so avoid chasing perfect combinations created by a handful of trades.
Practical application
Strategy analysis is most useful when every trade is labelled consistently. If the same setup appears under several names or its definition changes halfway through the sample, the comparison becomes unreliable.
Review checklist
- Keep a written definition for every strategy.
- Do not rename setups without normalising historical data.
- Compare return, risk and execution metrics together.
- Separate strategy versions when important rules change.
Frequently asked questions
Can I compare strategies with different frequencies?
Yes, but do not rely only on total profit. A higher-frequency strategy has more opportunities to accumulate P&L, so use per-trade and risk-adjusted metrics.
When has a strategy stopped working?
Look for persistent deterioration versus its historical behaviour with enough data and comparable conditions. Avoid making that conclusion from a short losing streak.