Tilt and revenge trading: catch it in the data, not in hindsight

What tilt is, mechanically

Tilt is the state in which a loss has changed how you evaluate the next trade. The setup quality bar drops, the time between decision and click shrinks, and size goes up because the goal has quietly changed from take good trades to get back to even. It is not a character flaw; it is how most people respond to an unexpected loss, and it is why casinos are carpeted and windowless.

Because the state changes behaviour in specific ways, the behaviour leaves specific marks: entries closer together than your normal rhythm, entries shortly after a losing exit, risk larger than the trade before, and — the tell — a sequence of these rather than one.

The signature in your journal

Two fields find most of it: the time of each entry and the risk taken on it. Mark every trade that was entered within a cooldown window of a losing close — thirty minutes is a reasonable default — and note whether its risk was larger than the loser's. Then add up the R-results of those flagged trades separately from the rest.

10:42  short XAUUSD   risk 1.0%   → −1R at 10:58
11:04  long  XAUUSD   risk 1.5%   → entered 6 min after the loss, risk ×1.5   → −1R
11:11  short XAUUSD   risk 2.0%   → entered 3 min after the loss, risk ×1.33  → −1R
Three trades, 4.5% of the account, 29 minutes. The first was a trade. The other two were tilt.

Sage's Psyche runs this detection automatically from the trade timestamps and risk already in your journal — no self-reporting — and shows the flagged entries with what they cost. Sage's automatic revenge-trade detection is one of the few places in trading where the honest number arrives without your cooperation.

Cost it in R

Feelings about tilt are vague and easy to argue with. A number is not. Take the flagged trades for a month and total their R. Most traders who do this for the first time find that tilt trades are a small fraction of their entries and a large fraction of their losses — and that removing them alone would have turned a losing month into a flat or winning one. That is the case for a rule, made by your own data.

Rules that hold when you are angry

"Don't revenge trade" is not a rule; it is an aspiration. A rule is something a machine could check. Three that work:

• A cooldown after any loss. No new entry for N minutes after a losing close. Thirty is a common start; the right N is whatever your flagged trades say. The point is to put time between the emotion and the click.
• A maximum number of trades per day. Tilt is a sequence. If your plan has two or three good setups a day in it, a hard cap of three trades ends the sequence before the third revenge entry exists.
• A daily loss stop, in R. Down 2R for the day, done for the day. Not "I'll be careful" — done. Under prop-firm rules this fence should sit well inside the firm's daily limit (the rules explained).

The check-in: know the day before it starts

Some bad days are visible at 8 a.m. Thirty seconds before the first trade — how did you sleep, how is your focus, how is your mood, is there a distraction today — gives you a baseline. Over a couple of months of check-ins you can compare your results on low-score days against high-score days. If the difference is stark, the rule writes itself: on a low-score morning, trade half size or not at all. That is not weakness; it is knowing your own numbers.

What the review is for

Do not review tilt trades by reliving them. Review them by category once a week: how many were flagged, what they cost, which rule would have prevented them, and whether the rule you already have was kept. If a rule was broken, the question is not "why was I weak" but "what made breaking it possible" — the trade was still available, the size was still adjustable, the day was still open. Change that, and the next tilt has nowhere to go.