2026-04-22 · 6 min read
Session edge, AI coach verdict, and running P&L for 2026-04-22.
Session slice (heuristic read)
Trim Review
How much of today's result was carried by a few outliers? The table recomputes the core session stats with the biggest movers removed. Trimmed rows are colored vs baseline: green = holds most of the move, amber = meaningful erosion, red = collapses or flips sign.
| Scenario | Trades | Net P&L | Total R | E[R] / trade | $ WR | PF |
|---|---|---|---|---|---|---|
| Baseline (all trades) | 71 | −$2336.78 | -4.47R | -0.063R | 53.5% | 0.65 |
| − single best + single worst | 69 | −$1561.13 | -2.98R | -0.043R | 53.6% | 0.71 |
| − top 4 + bottom 4 (≈5% each tail) | 63 | −$629.57 | -1.20R | -0.019R | 54.0% | 0.81 |
Trimmed trades: Single best + single worst — removed AKAN +$538.18 (best) · AGPU −$1313.83 (worst). Top 4 + bottom 4 — removed 4 from each tail (by $ P&L).
Bottom-heavy red session: most of the loss came from a few prints. The broader distribution is closer to neutral than the total P&L suggests.
Session verdict
Blind from the trade list alone, I would predict a mildly green or flat session: 53.5% winners, decent activity, and only the four trades at or above 1R look like a handful of tail events rather than a day-long bleed. The actual -$2336.78 means this session breaks, rather than extends, the recent coherent improvement in quality: the 7d window was running 0.158R per fill versus 0.013R in the prior-23d, with PF 1.71 versus 1.06, and today printed -0.063R with PF 0.65.
The geometry was concentrated, not broad; three downside tails, especially the two losses around 9:28am PT and 9:43am PT, did most of the damage while the rest of the book looks closer to noisy churn than systemic failure. What did not happen matters here: there was no outsized winner to offset those tails, with 0 trades at or above +2R and only one winner above +1R, so the session lacked the positive convexity that has been powering the last 7 days.
Path-wise this does not read like all-day forcing, because several windows earlier in the session paid and there was still a later +1R winner; it reads more like selective tail exposure without the matching home-run participation. Restraint helped in one sense because the damage was not spread across many material losers, but restraint also seems to have shown up as an absence of larger winners, so the miss was less about constant bad clicking and more about an imbalanced distribution.
Significance and conceptual math
The primary evidence is still the short-horizon trend, and it had been genuinely improving: versus the prior 23 days, the last 7 days showed win rate essentially flat at 45.4% versus 46.2% for a small -0.8 point change, but E[R] jumped by 0.145R, net dollars per day improved by $3364.63, and profit factor rose by 0.65 to 1.71. That is why the pack calls it coherent improvement, and today should be read against that backdrop.
Today then looks like a wobble or break inside that stronger 7-day upswing: win rate came in above both the 7d and baseline at 53.5%, but payoff quality collapsed, which is why P&L was negative despite more winners than losers. The single-day win-rate contrast versus baseline, z 1.372 with p about 0.1701, says only that today's hit rate was not unusually strong in a way the numbers can prove; the real issue was what each win and loss was worth.
Your average winner was $114.86 and average loser was -$203.08, so you needed 63.9% wins just to break even in today's payoff mix and only got 53.5%. By contrast, the 7d window needed only 32.7% wins because average winners were so much larger than losers, and the realized 7d win rate of 45.4% easily cleared that bar.
So the recent edge has been coming from payoff asymmetry, not accuracy, and today the asymmetry flipped against you. There is some low-power risk because the 7d comparison is only 5 session-days, but 282 fills is enough to take the directional message seriously.
The +5 percentage-point win-rate toy is worth $1128.68 for this exact session, which is meaningful but still not enough to fully erase a day driven by a few large losses and a lack of large winners.
What you might have missed
The counterintuitive part is that this was not mainly an overtrading day even with 71 positions. Most of the loss can be explained by the absence of positive tails: you had just one +1R winner, zero +2R winners, and three downside tails, so the book was missing the right tail that has recently made a sub-50% win rate profitable.
Also notable is that the ugliest damage clustered before 10:00am PT even though multiple earlier windows were net positive, which suggests the problem was not all early trading but a very narrow pocket where upside participation disappeared while downside tails still got through.
One thing to try next session
Next session, put a temporary rule on the opening phase: after the first loss worse than 1R before 10:00am PT, you cannot take another new position until you have seen one trade idea that can plausibly be held past 1 minute with at least 1.5R upside. The goal is to force one right-tail candidate into the book before stacking more short-hold attempts, because today's damage came less from too few winners and more from not capturing any meaningful upside tail.