The five-trade rule: how to know if a strategy actually has edge

Five winning trades in a row feels like proof. Three losing trades in a row feels like a death spiral. Both feelings are wrong by orders of magnitude — the statistics of small samples in trading are deeply counterintuitive, and getting them wrong is one of the largest single sources of bad decisions in retail trading.

Here's the actual math, the threshold of trades that actually means something, and how to journal so you can know.

The math behind it

If your true win rate is exactly 50%, what's the chance of going 5-0 in your next 5 trades? About 3.1%. That sounds rare. Now consider: across a year, most traders see dozens of 5-trade windows. The chance that AT LEAST ONE 5-trade window is 5-0 is much higher than 3%.

In other words: a 5-0 streak DOESN'T mean you have edge. It means you took 5 trades during a streak that would have happened naturally for a 50/50 trader.

The 95% confidence interval gets brutal

If you measure 60% win rate from a 20-trade sample, the 95% confidence interval is roughly 36-81%. That's a 45-point range. Your "60% win rate" is statistically indistinguishable from "the same as random."

Compare:

• 20 trades, 60% WR: 95% CI ~36-81%. Meaningless.
• 50 trades, 60% WR: 95% CI ~45-74%. Mildly informative.
• 100 trades, 60% WR: 95% CI ~50-69%. Clearly profitable.
• 200 trades, 60% WR: 95% CI ~53-67%. Real edge.

What this means for strategy testing

If you've tried a new setup 10 times and it's 7-3, you don't have evidence the setup works. You have an interesting hint that warrants taking the setup 40 more times before deciding. Most retail traders treat 10 trades as final evidence — both ways. They abandon a 4-6 setup that might be profitable, and they commit aggressively to a 7-3 setup that might be random.

Practical implications

• Don't size up after a winning streak. A 5-trade win streak is consistent with a 50% trader having an okay week. Sizing up based on it adds risk without evidence.
• Don't abandon a setup after a losing streak. A 4-trade loss streak is consistent with a 60% WR setup having a bad week. Wait for 50+ samples before retiring a setup.
• Track sample size alongside win rate. Win rate without sample size is noise. Tell yourself "my breakout-pullback is 64% over 73 trades" — that's informative. "My breakout-pullback is 64%" is not.

How to journal for sample-size awareness

Every setup-level analytic in your journal should show both the metric (win rate, average R, profit factor) AND the sample size. If a setup has fewer than 30 trades behind it, the metric should be displayed with a "limited sample" warning so you don't over-anchor on it.

TradeFlow Quantum's setup breakdown shows sample size prominently — you'll see something like "breakout-pullback: 124 trades, 71% WR, +1.42R avg" so the sample size is part of how you read the number. Setups with under 20 trades render dimmed.

The real five-trade rule

If you take exactly 5 trades and they're 5-0 or 0-5: you have learned nothing. You took 5 trades. Take 45 more before you adjust your strategy or your conviction.