Cutting a winning intraday trade early turned +41 into −2.6

By Strota Newsroom · 2026-08-24 · How Strota reports

The finding is blunt: simulating an early exit on a system designed to hold until the close took its total result from +41 R down to −2.6 R, and the win rate from 58% to 28%. That is not underperformance; that is the difference between an edge and no edge. The same study found fixed take-profit targets also lose to simply holding, and it identified which trade attribute actually carried the profits: medium-sized opening gaps. For anyone who has ever grabbed a small win out of fear and watched the trade they abandoned run on without them, the mechanism here will feel familiar.

Some notation first, because the unit matters. Results are measured in R — multiples of the risk taken per trade — so +41 R means the book gained forty-one times a single unit of risk across the sample. The sample was 237 trades over four sessions in mid-June, scored at end of day. One caution before any of the numbers lands: these are candidate changes examined on paper, none implemented, and the caveats at the end are load-bearing.

Start with the exit question, because it produced the sharpest result. About 90% of these trades already exit at the 15:30 square-off, the forced close of an intraday position. The entry alerts offer no priced exit signal — fade and reversal alerts never fired once on intraday entries during the window — so the study asked what would happen if trades were instead cut when the stock stopped being in play, using the last alert re-emission as a price proxy. Total R went +41 → −2.6. Win rate went 58% → 28%. The average cut came roughly six minutes after entry, before winners had time to develop. On the sample's one strong trend day, this policy collapsed a +30.7 R session to approximately zero.

Fixed take-profits fared better but still lost. Capping winners at +1 R produced +33.9 for the book; +1.5 R produced +36.7; even +2 R only reached +39.2 — against the +41.3 hold-to-close baseline. The pattern is mechanical rather than mysterious: capping every trade at a fixed gain amputates exactly the right tail that pays for everything else. The study measured the give-back directly — the average trade peaked at +0.51 R at best but captured only +0.17 R, roughly 0.34 R surrendered per trade — and still concluded holding wins, because what looks like generosity to the market is actually rent paid for access to the occasional outsized winner. The one exit change that might improve on holding is a trailing stop, which cannot be tested from end-of-day data alone.

There is one nuance where acting early does help: not entering at all. When a stock drops out of contention before its entry trigger fires — 46 such instances in the sample — respecting that signal means correctly avoiding a chase, at zero cost. Distinguishing pre-entry avoidance from post-entry meddling turns out to be the whole game.

The second half of the study examined which entries deserved the hold, and the gap data answered. Gap-and-hold entries — taking a position in the direction of the opening gap, betting the move continues — made +0.22 R per trade (about +38 R total) versus +0.05 R with no gap present. Gaps were the edge. But size decided everything. Small gaps under 1.5% earned +0.16 R. Medium gaps between 1.5% and 3% earned +0.30 R with a 64% win rate — about +29.8 R, effectively carrying the entire book. Large gaps of 3% or more lost −0.06 R with a 29% win rate: exhaustion, the same failure mode as chasing. Volume added a twist against intuition — gaps without climactic volume won 72% of the time at +0.42 R, while gaps on more than fifteen times normal volume fell to a 48% win rate, though that bucket is thin. Regime mattered too: bullish-tape entries made +0.27 R over 136 trades while bearish ones scraped −0.01 R. And position mattered most starkly of all — entries not pressing the day's high or low went 0-for-5.

Out of all this came one unimplemented candidate filter: take medium-sized gap-ups, in a bullish regime, pressing the day's high, without blow-off volume — and hold to the close. Skip large gaps, against-regime gaps, gap-down shorts, and anything not at the extreme.

Now the honest limits. Four days is a tiny window, three of them bullish-skewed, and a single trend session accounts for roughly 75% of the profit — remove that day and the picture dims dramatically. The juiciest buckets are thin: large gaps number 14 trades, extreme volume 29, gap-down shorts 34, the not-at-extreme bucket just 5. What deserves most trust is narrow and specific: the medium-versus-large gap distinction, and the finding that holding to EOD beats every early exit tried. Everything else awaits a fuller week of data.

More money stories

• One trading day was enough to flip four supposed edges
• First live test of our best long setup: profits held up, the 80% win rate fell to 55%
• Ask-heavy order books predict winning longs 66% of the time across 329 trades — but the short mirror fails
• Tested on fresh data, our high-win-rate setup wins less often but pays more
• Costs, not gross returns, decided which trades to skip: a sector filter earned more on fewer trades
• Big gaps don't sustain: Indian stocks that open far from the last close mostly hand it back