# How to Stop Revenge Trading Using Your Own Data

> How to stop revenge trading. Find it in your journal (quick re-entries after a loss, size creep), measure what it costs in R, then add rules that interrupt it.

Canonical: https://www.journalx.io/blog/how-to-stop-revenge-trading

To stop revenge trading, first find it in your own trades: quick re-entries after a loss, size that creeps up during a red day, and trades with no plan. Measure what those trades cost in R. Then add rules that interrupt the pattern before the next order: a cooldown after losses, fixed size, a written plan and a daily stop.

How to stop revenge trading is a question most traders ask the evening after a day that got away from them. The usual advice is to "stay calm" and "accept losses," which is true and almost useless at 9:52 a.m. with a red P\&L. A better approach is to treat revenge trading as a pattern in your data: it leaves fingerprints in your journal, you can put a number on it, and you can write rules aimed at the exact moment it starts.

## What revenge trading is, and why it happens

Revenge trading is taking a trade to win back a loss rather than because your [setup](/glossary/setup) is there. It rarely looks dramatic from the inside. It feels like spotting a decent entry a bit sooner than usual, with a bit more size, because you want the day back to flat.

That pull is well documented. Richard Thaler and Eric Johnson called it the [break-even effect](https://doi.org/10.1287/mnsc.36.6.643): after a loss, people become more willing to take a gamble that offers a chance to get back to even. Professionals are not immune. In a study of Chicago Board of Trade proprietary traders, Joshua Coval and Tyler Shumway found that traders regularly took above-average risk in the afternoon to recover morning losses, and that the losing traders paid higher prices to buy and accepted lower prices to sell than before ([Journal of Finance, 2005](https://doi.org/10.1111/j.1540-6261.2005.00723.x)). In other words, the urge to get it back also makes the fills worse.

## How revenge trading shows up in your trade data

You don't have to guess whether you revenge trade. Sort your trades by entry time and add three columns:

1. **Minutes since the last trade closed, and whether that trade lost.** Revenge trades cluster in the first few minutes after a losing close.
2. **Size versus plan.** Divide each trade's dollar risk by your normal [risk per trade](/glossary/risk-per-trade). A value of 1.5 means you risked 50% more than planned. Revenge trades tend to drift above 1.0 as the day gets worse.
3. **Setup or plan recorded.** A trade with no setup tag and no written stop is often a reaction, not a decision.

Then compare [expectancy](/glossary/expectancy) in R for the flagged trades against the rest. Here is what that looks like in one illustrative journal of 240 day trades, split by what happened just before each entry.

![Expectancy by what happened just before the trade, across 240 day trades: +0.26R after a win or on the first trade of the day (110 trades), +0.17R after a loss when the trader waited 10 minutes or more (80 trades), −0.25R within 10 minutes of one loss (34 trades) and −0.66R within 10 minutes of two or more losses (16 trades).](https://assets.journalx.io/marketing/blog/how-to-stop-revenge-trading/expectancy-after-a-loss.avif)

The 190 trades taken after a win, or after a loss with a real pause, made 42.2R together. The 50 trades entered within 10 minutes of a loss lost 19.1R, giving back about 45% of everything else. The strategy itself was fine. Waiting ten minutes after a loss still produced +0.17R per trade, so the problem was never "trading after losses." It was trading *straight* after them.

Ten minutes is a starting point, not a law. A scalper might find the damage in the first two minutes, and a swing trader might look at "same day as a loss" instead. Try a few gaps and see where expectancy turns negative. A quick re-entry on a valid, planned setup is not revenge, which is why the setup column matters.

## What one revenge trading morning costs

Averages hide how fast it happens. This is a single morning from the same kind of journal: planned risk of 1R per trade and a daily loss limit of 3R.

![One bad morning: the first two trades at planned size lose 1R each, then four more trades follow within one to two minutes of each close at 1.5x, 2x, 2x and 2.5x planned size. The day's P\&L falls from −2R to −7.2R, more than twice the 3R daily limit.](https://assets.journalx.io/marketing/blog/how-to-stop-revenge-trading/one-bad-morning.avif)

The first two trades were ordinary losses at planned size, the kind any strategy produces (our post on [how long a good strategy can keep losing](/blog/losing-streaks) shows how often). Everything after that was revenge. The gaps between trades shrank to a minute or two, size climbed to 2.5 times the plan, and the day ended at −7.2R instead of −2R. Two normal losses cost 2R. The four trades that followed cost 5.2R, which at $200 per R is $1,040 of damage that the strategy never asked for.

## How to stop revenge trading: rules that interrupt it

Willpower is weakest right after a loss, so the fix has to be decided earlier. Each rule below targets one of the fingerprints from the data.

### 1. Set a cooldown after every loss

No new order for a fixed time after a losing trade closes. Pick the length from your own numbers, then fill it with something useful: stand up, and write the journal note for the losing trade before you look at a chart again. Writing the note doubles as the timer, and it makes you describe the loss in words, which is hard to do while chasing it.

### 2. Fix your size before the session

Write your dollar risk per trade before the open and treat it as a ceiling for the day. Size can go down after losses if your plan says so, but never up on a red day. With size held at 1R, the morning above would have ended near −4.6R instead of −7.2R, and it removes the size-creep column from your data entirely. If you need a refresher on the math, see [position sizing](/glossary/position-sizing).

### 3. No written plan, no trade

Before every order, write the setup, entry, stop, target and size. A revenge trade rarely survives having to name its setup and its stop. Our [pre-trade checklist](/blog/pre-trade-checklist) has a version you can copy, and a [pre-trade plan](/glossary/pre-trade-plan) also gives you the planned 1R that makes the size-versus-plan column possible.

### 4. Stop after a set number of losses, and at a set loss

A loss-count rule (for example, two losses in a row means a 30-minute break) catches the spiral early. A [max daily loss](/glossary/max-daily-loss) ends it. In the example, a 3R limit would have ended the day at −3.5R, half of what it became. [How to set a daily loss limit](/guides/how-to-set-a-daily-loss-limit) walks through sizing the limit from your win rate and what to do the moment it hits.

### 5. Write the rule as "if this, then that"

"If I take a loss, then I close the order ticket, write the note and wait 15 minutes" works better than "don't revenge trade." Plans in this if-then form had a medium-to-large effect on whether people followed through on their goals in [a meta-analysis by Gollwitzer and Sheeran](https://doi.org/10.1016/S0065-2601%2806%2938002-1). Put the rule in your [trading plan](/glossary/trading-plan) where you will see it before the open.

## Check every week that the rules are working

Rules only work if you check them. Once a week, count the trades entered inside your cooldown, the trades above planned size, and the trades with no plan, and add up what they cost in R. Tagging them with a mistake tag as you go makes this a two-minute job (see [how to tag trades](/guides/how-to-tag-trades)). You are looking for two numbers to fall: how many revenge trades you took, and their total cost.

Expect the count to drop before it reaches zero. A week with three revenge trades instead of twelve is real progress, and the R saved is the clearest reason to keep going.

JournalX lets you plan each trade before you take it and records your risk in R, so you can tag the trades that broke your rules and compare their expectancy with the rest of your journal.
