@quantpad: Gap fill strategy backtest on 7,744 S&P and Nasdaq futures opens since 2011. "Wider gaps make for better gap fill reversal trades." That is Wysetrade, in a video with 375K views. I tested the claim rather than proposing a gap strategy: every 09:30 open on ES and NQ, 2011 to November 2026, against the same contract's prior 15:59 close. How often the gap filled by the close, ES then NQ: under 0.1%, 91% and 94%. 0.1 to 0.25%, 76% and 82%. 0.25 to 0.5%, 54% and 62%. 0.5 to 1%, 32% and 45%. Over 1%, 21% and 24%. Wider gaps fill less. A level the same distance the other way got touched almost as often: 59.7% vs 57.5% on ES, 62.1% vs 58.6% on NQ. Fading after the first 5-minute bar closes against the gap, target the prior close, stop one gap away, one contract, $16.50 costs: gaps over 1% made +$69.89 a trade on ES (202 trades) and lost $205.20 on NQ (297 trades). The 95% block-bootstrap range of widest minus narrowest crosses zero on both. Walked forward (expanding yearly from 2014, 13 folds, minimum gap and stop picked on past years only): ES +$7,853, NQ -$44,483. Doubling costs: +$53.39 and -$221.70 a trade. Buy and hold one contract: ES +$312,757, NQ +$488,754. The confirmation bar matters: across all gaps, every 5-minute setting lost on NQ, every 1- and 15-minute setting made money. On 11 big stocks over the last 365 days (one venue), 7 filled less often than the other side. Not modelled: his discretionary zones and trend reads. My mechanical version of his key level and Bollinger filters fired 10 times in 16 years, too few to say anything. Sessions with halted minutes are excluded, so the most violent days are missing. Read a gap as a distance, not a magnet. Past half a percent, do not count on the fill. I built this in QuantPad. It has every trade on the CME since 2010. Link in bio. #QuantPad #backtesting #futurestrading #tradingstrategy #algotrading

QuantPad

Open In TikTok:

Region: US

Thursday 05 November 2026 19:02:19 GMT