Revenge Trading: How to Find It in Your Own Data

Revenge trading is obvious afterwards and invisible while it happens. Here is what it looks like in trade data, and how to catch the pattern before it costs the account.

Daniël Vermes

Tradeflow Editorial Team

Aug 10, 2026

10 min read

Revenge trading is the pattern of entering a new position primarily to recover a recent loss rather than because a valid setup appeared. It typically shows up as increased position size, shortened time between trades, and declining setup quality in the minutes or hours following a loss. It is the most commonly cited cause of account destruction in retail trading, and the most consistently misdiagnosed.

Everyone knows what it is. Almost nobody can tell you when they last did it.

That gap is the entire problem, and it is why the standard advice does not work.

Why "just take a break" fails

The usual guidance is to step away after a loss. Breathe. Walk around. Come back when you are calm.

It fails for a simple reason: it requires you to notice you are tilted, and the defining feature of the state is that it does not feel like tilt from inside. It feels like conviction.

You do not experience yourself entering a bad trade out of frustration. You experience yourself spotting an opportunity, seeing it clearly, and acting decisively while others hesitate. The urgency feels like edge. The size feels justified because the setup is obviously good. Every internal signal you would normally rely on has been recruited to support the decision.

Advice that depends on self awareness in the exact moment self awareness is compromised is not a strategy. It is a wish.

What works instead is detection after the fact, repeated often enough that the pattern becomes visible, followed by a mechanical rule that does not require you to be calm to apply it.

What it actually looks like in the data

Revenge trading has a signature. It is remarkably consistent across traders and it is visible in trade history even when it was completely invisible at the time.

Position size increases after a loss. The clearest marker. Not a doubling, usually. A 40 or 60 percent step up, justified at the time by conviction rather than by the setup. Look at your risk per trade in the three trades following your worst losses and compare it to your baseline.

The gap between trades shortens. If your normal spacing is forty minutes and the trade after a loss came in six, that is not a coincidence. Time to next entry is one of the most reliable indicators in any trading history and almost nobody measures it.

Setup quality drops. Trades that do not match any of your named patterns start appearing, or your loosest setup starts showing up disproportionately. If a meaningful share of your trades are untagged or tagged as "other", check when they occurred.

Instrument drift. You lose on your primary instrument and take the next trade somewhere unfamiliar. Often on something more volatile, because it offers a faster route back to flat.

Clustering. Losses arrive in tight groups rather than spread across the session. Three or four in ninety minutes is rarely four independent decisions. It is one decision repeated.

Any one of these in isolation might be nothing. Two or three together, appearing repeatedly after losses, is a pattern rather than a bad day.

The part that makes it expensive

Revenge trading would be manageable if it always lost. It does not, and that is what makes it durable.

Sometimes the oversized trade taken in frustration works. You recover the loss, finish the day green, and the entire episode gets recorded in memory as decisive trading under pressure. The behaviour was rewarded, so it is now more likely to recur, and the version of you that shows up after the next loss is slightly more confident about it.

This is why revenge trading is not really a psychology problem. It is a reinforcement problem. The behaviour is intermittently rewarded, which is the schedule that produces the most persistent habits of any known reinforcement pattern. It is the same structure that makes slot machines work.

Which has a practical consequence: reviewing only your losing trades will hide the problem. The profitable revenge trades are the ones building the habit, and they are invisible in any review organised around P&L.

How to actually find it

Not from memory. Your recollection of your own trading is reconstructed rather than recorded, and it reconstructs in your favour without asking permission.

Do this instead, on synced trade data covering at least three months.

Find your ten worst losses. Not by percentage, by absolute impact on the account.

Look at the three trades after each one. For each, note position size relative to your average, minutes elapsed since the previous exit, and whether the setup matches one of your named patterns.

Count how many of the thirty show two or more markers. If it is more than about six, revenge trading is a live pattern in your account rather than an occasional lapse.

Then look at how those thirty trades performed in aggregate. This is the number that changes behaviour. Most traders find a meaningfully negative expectancy concentrated in a small number of trades taken within an hour of a loss, which means a single mechanical rule would have recovered a large share of their annual losses.

That figure, in your own currency, is more persuasive than any amount of advice about breathing.

Doing this without losing a weekend

Every step above is filtering and cross referencing. By hand, across a few hundred trades, it is an afternoon of work, which is why it does not get done and why most traders never see their own number.

This is where Tradeflow is useful in a way a spreadsheet is not. You ask directly rather than building filters. Show me my ten largest losses. Show me the trades that came within an hour of a loss. Tag those as tilt. Group by that tag and show me the expectancy.

It reads your synced history, applies the tags across every matching trade, and returns the answer. Retroactively, so a suspicion you form on a Sunday evening can be tested against three months of history in one request instead of an evening of scrolling.

That retroactive part matters more than it sounds for this specific problem. Revenge trading is only visible in aggregate. One instance is a bad trade. Thirty instances with a shared signature and a measurable cost is a pattern you cannot argue with, and getting to that view is the whole battle.

Conversational AI is included on every plan from $19 a month.

The rules that actually work

Once you have your number, the fix is mechanical rather than emotional. Rules that require no judgment at the moment of application, because judgment is the thing that is compromised.

A time lock after a threshold loss. No new position for thirty minutes after any loss over 1R. Not a suggestion, a rule. The specific duration matters less than that it is fixed in advance and not negotiable in the moment.

A daily loss limit that ends the session. Two losses, or a set percentage, and you are done for the day regardless of what the chart is doing. Most traders resist this because it occasionally costs them a good afternoon. Check your data on what those afternoons actually returned before deciding it is too strict.

No size increases within the session. Position size set before the open and not adjustable during. This removes the single most damaging marker entirely.

Same instruments only. If it is not on your list for the day, it is not a trade.

Log your state before entry, not after. Two words, before you know the outcome. This is the only field that lets you measure the problem going forward, and it takes about three seconds.

Track compliance with whichever rule you pick as a number, weekly, the way you would track any other metric. That figure, not your P&L, is what you are trying to improve. We covered the review process in the 20 minute weekly trading review.

What to do in the moment

Since you will not always catch it early, one thing worth having.

Write your rule somewhere you will physically see it before you can place an order. Not in a document. On the monitor. The moment you are trying to interrupt is one where your reasoning has already been recruited to the other side, so the intervention has to be something that does not require reasoning to work.

If you find yourself constructing an argument for why this particular situation is an exception, that argument is the symptom. Genuine setups do not require a case for the defence.

Key takeaways

Revenge trading does not feel like tilt from inside, which is why advice depending on self awareness fails. Its signature is measurable: larger size, shorter gaps between trades, degraded setup quality, instrument drift, clustered losses. Profitable revenge trades are the dangerous ones, because intermittent reward produces the most persistent habits. Reviewing only losses hides the pattern, since the reinforcing instances were winners. Find it by examining the three trades following each of your ten worst losses, then measuring their aggregate expectancy. The fix is mechanical, not emotional. Fixed rules applied without judgment, because judgment is what is compromised. Track rule compliance as a weekly number, the same way you track any other metric.

Frequently asked questions

What is revenge trading?

Entering a position primarily to recover a recent loss rather than because a valid setup appeared. It usually shows as increased size, shortened time between entries, and lower quality setups in the period following a loss.

How do I know if I am revenge trading?

Not reliably in the moment, which is the core difficulty. Check it afterwards in your data: look at the three trades following each of your largest losses and compare position size and time between trades against your baseline. The pattern is obvious in aggregate and invisible individually.

Why does revenge trading feel like conviction?

Because the emotional state recruits your reasoning rather than overriding it. You do not experience frustration, you experience clarity and urgency. Every signal you would normally use to check yourself is already supporting the decision.

Does revenge trading always lose money?

No, and that is precisely why it persists. Intermittently rewarded behaviour is the most durable kind. The revenge trades that worked are the ones reinforcing the habit, and they are invisible in any review organised around losses.

What is the most effective rule against it?

A fixed time lock after a threshold loss, typically thirty minutes after any loss over 1R. It works because it requires no judgment at the moment of application, which is when judgment is least available.

How is revenge trading different from tilt?

Tilt is the broader emotional state of degraded decision making, which can follow wins as well as losses. Revenge trading is the specific behaviour of trading to recover, and it is one of the most common expressions of tilt.

How many trades do I need before the pattern is measurable?

Enough to include ten meaningful losses, which usually means at least three months of history. Below that you are looking at individual bad days rather than a pattern.

Find your number

Tradeflow syncs every trade automatically from 600+ brokers and platforms and lets you tag and analyse by asking rather than filtering. Ten minutes to find out what trading after a loss has actually cost you.

Start free for 7 days. From $19 a month afterwards, AI included on every plan.