@quantpad: How many trades a backtest needs before its win rate means anything, measured on 15 years of real Nasdaq and S&P 500 futures trades, 2011-2025. Casper Trading's video shows a month where "81% of those 16 trades were winning trades". We took his first candle rule (mark the 9:30-9:45 candle, buy back at the first 5-minute gap through it, stop at candle one, 2 to 1 target) and ran it on every session since 2011, then cut its trades into every stretch of 16, 30, 50, 100, 200 and 500. Nasdaq (NQ): 1,880 trades, 30.96% hit the target. Yet 90% of 16-trade stretches landed anywhere from 12.5% to 50%. The best did 10 of 16; none reached 81%. At 50 trades the middle 90% still spans 20% to 40%. The number: the win rate only settles within 5 points of the long-run rate at 206 trades on NQ and 217 on the S&P (ES, 1,855 trades, 32.18%). Hot streaks: after the hottest 10% of 16-trade stretches (47.9% on average), the next 16 trades won 28.65%. The correlation between one stretch and the next was -0.03. Baseline: a coin picking long or short at the same entries swings exactly as much, 12.5% to 50% over 16 trades. That spread is arithmetic, not this rule. Also checked: 2011-2018 won 30.24%, 2019-2025 31.74%. Win rate only; costs were tested on this rule in our earlier video (-$373 net on NQ over 15 years). Takeaway: get about 200 trades before you trust a win rate, not 16. Every CME futures trade since 2010 is built into QuantPad. I built this in QuantPad. Link in bio. #QuantPad #backtesting #daytrading #futurestrading #tradingstrategy

QuantPad

Open In TikTok:

Region: US

Sunday 01 November 2026 18:07:05 GMT