← Blog · · 6 min read
My Shortest Trades Lost the Most Money
I went back through 43 trades my bot actually took and sorted them by one thing I usually ignore: how long the position stayed open. The pattern was not subtle. The faster I got out, the worse I did.
I always assumed the entry was the thing to obsess over — the perfect signal, the perfect candle. So I audited the log expecting the winners and losers to split by which setup triggered them. They did not. They split by time in trade. Here is the exact breakdown from that sample.
The three time buckets
- Held under 10 minutes: 18% win rate, −$7.47 net. The single worst bucket. Most of these were positions I bailed on inside the first few minutes because the price did not move my way immediately.
- Held 10 to 30 minutes: 69% win rate, +$5.05 net. The sweet spot. Enough room for the trade to actually work, not so long that it turned into a hostage situation.
- Held 2 to 4 hours: 100% win rate, +$2.22 net. Small sample, but every one of them closed green. These were the trades I let breathe all the way to target.
Look at the under-10-minute row again. An 18% win rate is not a strategy, it is a coin that lands on its edge. And it bled $7.47 — more than the profitable buckets made — from trades that never got the chance to be right or wrong. I closed them on a feeling.
The uncomfortable part: my exits, not my entries
Here is what actually happened in that under-10-minute bucket. The entry signals were the same ones that made money in the 10-30 minute band. Same setups. The difference was that I panicked out early. A trade would go −0.3% in the first two minutes — which is noise, not information — and I would close it to "protect capital." Then it would turn around without me.
That is the expensive habit. A small early drawdown is not a signal. It is the normal texture of a position finding its footing. When I sorted the closes by reason, the manual panic-close on a tiny loss was one of the most reliably unprofitable things I did. The mechanical exits — let it hit the stop, or let it hit target — did far better than my judgement in the moment.
Why fast trading is a fee machine too
There is a second tax hiding in that top bucket, and it has nothing to do with being right. Every one of those sub-10-minute trades paid a full round trip of fees. Taker in, taker out. Do that 30 times a day and the fees alone can eat more than any edge a quick scalp ever had.
I have written before about how fees quietly kill scalping, and the hold-time data is the same story from a different angle: the shorter the average trade, the more times you pay the house, and the less room each trade has to overcome that cost. You can put real numbers on your own round-trip drag with the exchange fee comparison calculator and the break-even calculator — the move you need just to get back to zero is bigger than most fast traders think.
What the winners had in common
The best exits in the whole log were not clever. They were the trades that reached their take-profit or got trailed out by a stop that only moved in my favour. When I filtered for positions that actually hit target, the win rate was above 90% and they carried the account. The common thread was simply that I stopped touching them. I set the stop, set the target, and let the trade use the 10 to 30 minutes it needed.
That is the whole lesson, and it is boring: give the trade time to work, and size it so you can afford to. The reason I could sit through a −0.3% wobble on the winners is that the position was small enough that the wobble did not scare me. On the panic closes, it was usually too big — so a normal drawdown felt like an emergency. Position size and patience are the same problem wearing two hats. The position-size calculator fixes the first half; discipline fixes the second.
How I trade it now
- No manual close inside the first 5 minutes unless there is real news. A tiny early loss is noise. I let the stop do its job.
- Size so a normal drawdown is survivable. If a −1% move against me feels like a crisis, the position is too big — check the max drawdown you can take at your leverage with the max-safe-leverage calculator.
- Let winners reach target. The 100% bucket exists because I stopped harvesting green trades early. Cutting a winner at +$0.20 to "lock it in" is how you end up with an average win too small to cover the losers.
The takeaway
I spent months tuning entries. The audit says the entries were fine — my clock was the problem. The trades I strangled in the first ten minutes lost more than the patient ones ever made. If you are an active trader and you have never sorted your own log by hold time, do it. There is a decent chance your best edge is simply to stop closing trades before they have a chance to work.
Run your own numbers first: break-even after fees, position size, and profit after costs. New to any of these? Start with the free trading academy.
Educational only — not financial advice. Figures are from my own paper and live trade logs (n=43) and are not a promise of future results. Leverage can lose your entire deposit.