Waiting for a pullback before entering a breakout

Fifty retracement depths were swept, from one tick to fifty, with the bracket re-anchored to the actual fill. None beat the baseline. The best was one tick deep at −$3,840; the worst gave up $50,125.

The price effect is real and strongly positive — waiting is worth between $7.80 and $217 a trade on the sessions where the entry fills, with t as high as +8.5. It is outweighed every time by which sessions get declined. At every depth tested, 100% of the declined sessions were winners at baseline, worth seven to eight times the average trade.

That is mechanical rather than unlucky. A session that never retraces is a session that left without you, and on a continuation strategy those are the target hits. Any entry filter conditioned on an adverse move removes only the right tail.

The mirror — entering after a move in favour, which declines 100% losers — looked like the survivor and did not replicate. Held at the depth chosen on the first strategy and run on three others it read −$14,946, −$12,981 and −$7,974 against the original's +$5,144, pooling to t = −0.38. A filter can remove a population that is almost entirely losers and still lose, because the entry change is paid on everything it keeps.