The move that feels smartest in the moment is usually the one you should look at hardest. Selling half of a winner once it moves a full R in your favor is that move. Your stop comes to breakeven, the trade is now free, and the part of your brain that hates giving profits back gets exactly what it wants. I have done it more times than I can count, and for a while I thought it was just obviously correct. Then I actually ran the numbers on my own trade history and found that on one of my strategies it was quietly costing me money.
The reason is not complicated once you see it, but it is easy to miss because the thing that improves is the number everyone stares at. Scaling out raises your win rate. It almost has to. And a higher win rate feels like progress even when your total return is going the wrong way.
What scaling out actually does to the distribution
Think in R-multiples, where R is the amount you risked on the trade. A full loss is minus one R. If you take half off at plus one R and let the rest run with a trailing stop, you have split every trade into two smaller bets with different personalities.
The first half is now a coin-flip-ish scalp that mostly comes back green, because plus one R is close enough to your entry that price touches it often. That half pulls your win rate up and smooths your equity curve. It feels great. The second half is where the real story is, and it is a smaller position than it used to be, so every genuine home run now pays you less.
Here is the part that bites. Trend strategies do not make their money on the median trade. They make it on the handful of trades that go plus eight R, plus twelve R, the ones that run for weeks. When you sold half at plus one R, you were holding a half-size position through the exact move that was supposed to fund your whole year. You capped your right tail to buy a smoother middle. For a strategy whose entire expectancy lives in that right tail, that is a bad trade with yourself.
Mean-reversion is the mirror image. Those strategies do not have a fat right tail to protect. Price pokes into an extreme, snaps back, and the move is mostly done. There is rarely a plus twelve R hiding in a mean-reversion setup, because the whole thesis is that the move exhausts. Scaling out early there is not capping anything you were going to get anyway, and locking in the bounce before it fades can genuinely improve your results. Same mechanic, opposite verdict, entirely because of the shape of the payoff.
Scale-out ladders versus all-out trailing stops
Broadly there are two families of exit, and they suit different systems.
- Scale-out ladder. You pre-decide levels, say a third at plus one R, a third at plus two R, and let the last third ride a trail. Your average win comes down but your win rate and your consistency go up. This fits mean-reverting and range-bound systems, and it fits any strategy where the psychological cost of round-tripping a winner would make you abandon the plan.
- All-out trail. You hold the entire position and exit only when a trailing stop gets hit, whether that is a moving average, a chandelier stop off recent range, or a structural level. Your win rate drops, because more trades that were briefly green get stopped back near breakeven, but your winners are full size when they run. This is the natural fit for trend following, where you are explicitly paying for the right to catch the outlier.
The honest framing is that a scale-out ladder trades expectancy for smoothness, and an all-out trail trades smoothness for expectancy. Neither is free. If your edge is thin and your discipline is thinner, the smoother path that keeps you in the seat can be the higher-expectancy choice in practice even if it looks worse on paper, because the paper version assumes you actually hold through the drawdowns without flinching. Most people do not.
How to test this on your own trade history
You do not need a fancy simulator to answer this for yourself. You need your closed trades logged in R-multiples and a couple of hours. The whole point is to stop arguing from feel and let your own data settle it.
- Convert every trade to its R-multiple at the moment you exited, and record the maximum favorable excursion, meaning how far in your favor the trade went before you got out. This is the single most important column, because it tells you what was on the table.
- Reconstruct the counterfactual. For each historical trade, apply the alternative exit rule to the price path it actually took. Take half at plus one R here, or trail it there. You are replaying the same setups under a different exit and reading off the new R-multiple.
- Compare the full distributions, not just the averages. Look at win rate, average win, average loss, and expectancy per trade. Then look specifically at what happened to your top handful of trades under each rule. If scaling out chopped your three best trades in half, that is the cost, sitting right there.
- Watch the tail, not the middle. If your biggest winners barely move between the two rules, your strategy does not have much of a right tail to protect and scaling out is close to free. If they collapse, your edge lives in the tail and you should think hard before capping it.
One failure mode to name plainly. People backtest a scale-out ladder, see the win rate jump from roughly the low fifties to the high sixties, and conclude the rule is better. Win rate is not the objective. A rule can win more often and still hand you less money at the end of the year. Always carry the comparison through to total return or expectancy, because that is the number that pays rent.
A rule of thumb I actually use
Before I touch an exit rule I ask one question. Does this strategy make its money from the median trade or from the outliers. If most of the profit comes from a few big runners, I keep the position whole and manage it with a trail, and I make peace with a lower win rate and the round-trips that come with it. If the profit is spread evenly across many trades that behave similarly, or if the setups are mean-reverting and the move exhausts fast, a scale-out ladder usually helps and it makes the equity curve something I can actually live with.
The mistake is applying one exit philosophy to every strategy because it fits your temperament. Your temperament is real and worth respecting, but it belongs in the position-sizing conversation, not the exit-structure one. Match the exit to the payoff shape first. Then, if the correct exit is one you know you will not be able to hold through, size down until you can, rather than switching to a comfier exit that quietly bleeds your edge. The comfortable exit that you follow beats the optimal exit that you abandon, but only after you have checked, on your own numbers, how much that comfort is costing you.