Tracking the gap between trades, and the cluster of fast ones that lost
I added a column to my journal for the number of minutes since my previous trade. It was easy to calculate from the timestamps, and when I sorted the data, there was a clear pattern.
Trades that came within a few minutes of the previous one had a clearly worse result than those with a longer gap. Many of them were after a loss, and some were after a win. In both cases, I was acting on momentum or emotion, not on a fresh setup.
When I looked at specific examples, a lot of them were cases where I hadn't properly reset. I was still in the mood of the last trade, whether frustrated or excited, and I took the next signal without thinking. The gap between trades acted as a measure of how much I paused.
I put a rule in place: a minimum gap of ten minutes between trades unless I'm managing the same trade. It's an arbitrary number, but it's been enough to force a pause. The number of rushed trades went down, and my average quality went up a bit.
Tracking the gap between trades, and the cluster of fast ones that lost
Re: Tracking the gap between trades, and the cluster of fast ones that lost
I added the same column after reading something similar and saw the same pattern. The fast trades after a loss were the worst, but the ones after a win were not much better, which surprised me. I think after a win I felt confident and lowered my standards. The ten minute rule seems arbitrary, but I agree that the number matters less than the pause itself. Mine is fifteen, and during that time I write a short note on the last trade.
It’s Fairman 
Re: Tracking the gap between trades, and the cluster of fast ones that lost
Something I would check is whether the gap matters differently depending on the setup. For me, one setup naturally produces two signals close together, like a failed first entry and then a valid second one, and those were fine. The bad fast trades were the ones that were a different setup or no real setup at all. So I made an exception for the planned second entry and kept the rule for everything else. Being precise about which fast trades are the problem kept the rule from blocking good entries.
It’s Fairman 
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Tracking the gap between trades, and the cluster of fast ones that lost
Measuring the gap from the previous exit rather than the previous entry seems the cleaner version. A trade held for 25 minutes followed by a new entry 2 minutes after closing it is exactly the rushed case, and an entry to entry gap would score it as a long pause.
There is also another explanation worth ruling out before blaming mood. Fast follow up trades tend to happen in fast markets, so the gap could partly be standing in for volatility around data releases or the session open.
A quick split tests that. Suppose a sample of 200 trades has 40 opened within 5 minutes of the last exit, and 25 of those 40 come in the half hour after a release, then the loss pattern may belong to news conditions as much as to a lack of reset.
If the fast trades are bad across quiet hours too, the pause rule is clearly earning its keep. If they are only bad around news, a news window rule would block fewer good entries.
What goes into the short note mentioned above during the fifteen minutes? A fixed question, such as whether the last trade followed the plan, seems easier to keep up than a free comment.
There is also another explanation worth ruling out before blaming mood. Fast follow up trades tend to happen in fast markets, so the gap could partly be standing in for volatility around data releases or the session open.
A quick split tests that. Suppose a sample of 200 trades has 40 opened within 5 minutes of the last exit, and 25 of those 40 come in the half hour after a release, then the loss pattern may belong to news conditions as much as to a lack of reset.
If the fast trades are bad across quiet hours too, the pause rule is clearly earning its keep. If they are only bad around news, a news window rule would block fewer good entries.
What goes into the short note mentioned above during the fifteen minutes? A fixed question, such as whether the last trade followed the plan, seems easier to keep up than a free comment.
-
FXS Prop Desk
- Posts: 1689
- Joined: Wed Oct 07, 2026 9:52 am
Re: Tracking the gap between trades, and the cluster of fast ones that lost
Fast follow-up trades are often reactions to the previous trade rather than new setups.
A minimum gap rule, for example no new trade within ten minutes of the last one closing, is easy to follow and directly targets that cluster. It might be worth testing what the results would have been with such a rule applied to the last few months of data.
A minimum gap rule, for example no new trade within ten minutes of the last one closing, is easy to follow and directly targets that cluster. It might be worth testing what the results would have been with such a rule applied to the last few months of data.
FXS Official Team | The Idea Lab & Strategy
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Tracking the gap between trades, and the cluster of fast ones that lost
Testing the ten minute rule on past trades, as suggested above, has one trap. Once a fast trade is struck out, the trade after it should be measured from the exit before, because with the rule running live the struck trade never happened.
Take a made up cluster where trade A closes at 10:00, B opens at 10:02 and closes at 10:04, and C opens at 10:11. A simple filter on the recorded gaps drops both B and C, but live, B never exists, so C sits 11 minutes after the last real exit and goes ahead.
It looks tiny, but it decides how many trades the rule really removes. Rerunning the trades in time order with the rule applied, rather than filtering the recorded gaps, gives the honest count.
Wins belong in the same comparison. Say 40 fast trades average minus 0.25R while 160 others average plus 0.15R, then the rule lifts the total from 14R to 24R, though if some of those fast trades were the planned second entries mentioned above, keeping them could be worth more than a blanket pause.
Does the pause also apply after a manual close, or only after a stop or target? A trade closed early by hand is often the most emotional exit of all.
Take a made up cluster where trade A closes at 10:00, B opens at 10:02 and closes at 10:04, and C opens at 10:11. A simple filter on the recorded gaps drops both B and C, but live, B never exists, so C sits 11 minutes after the last real exit and goes ahead.
It looks tiny, but it decides how many trades the rule really removes. Rerunning the trades in time order with the rule applied, rather than filtering the recorded gaps, gives the honest count.
Wins belong in the same comparison. Say 40 fast trades average minus 0.25R while 160 others average plus 0.15R, then the rule lifts the total from 14R to 24R, though if some of those fast trades were the planned second entries mentioned above, keeping them could be worth more than a blanket pause.
Does the pause also apply after a manual close, or only after a stop or target? A trade closed early by hand is often the most emotional exit of all.
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Tracking the gap between trades, and the cluster of fast ones that lost
A strategy wins 8 of its first 10 trades. Excitement follows. The statistics say very little.
A rough rule: the uncertainty in a measured win rate shrinks with the square root of the number of trades. With 10 trades, a measured 80 percent could easily reflect a true rate between 50 and 95. With 100 trades, an observed 60 percent suggests a true rate somewhere between about 50 and 70. With 400, the band tightens to roughly 55 to 65.
For scalpers, hundreds of trades accumulate within weeks, so waiting for a decent sample is feasible.
The trap is evaluating changes too soon. Adjusting a strategy after 20 trades means chasing noise.
A discipline that helps is to define in advance how many trades will pass before any judgement, such as 150, and to resist modifications until then.
What sample size is required here before a strategy is declared good or bad?
A rough rule: the uncertainty in a measured win rate shrinks with the square root of the number of trades. With 10 trades, a measured 80 percent could easily reflect a true rate between 50 and 95. With 100 trades, an observed 60 percent suggests a true rate somewhere between about 50 and 70. With 400, the band tightens to roughly 55 to 65.
For scalpers, hundreds of trades accumulate within weeks, so waiting for a decent sample is feasible.
The trap is evaluating changes too soon. Adjusting a strategy after 20 trades means chasing noise.
A discipline that helps is to define in advance how many trades will pass before any judgement, such as 150, and to resist modifications until then.
What sample size is required here before a strategy is declared good or bad?