forex-scalping.com

Does Taking Half Off at 1R Help? Run the Numbers

Using the hit rates from the previous drill, 60 percent of trades reach 1R and 10 percent reach 3R, compare three exit plans. Plan A: close everything at 1R. Plan B: close half at 1R, move the stop to breakeven, and leave half for 3R. Plan C: close everything at 3R.

Plan A: 0.6 times 1 minus 0.4 times 1, which is plus 0.20R.

Plan C: 0.1 times 3 minus 0.9 times 1, which is minus 0.60R.

Plan B needs three outcomes. First, trades that never reach 1R, probability 0.4, lose 1R. Second, trades that reach 1R and then 3R, probability 0.1: half banks 1R and half banks 3R, so 0.5 plus 1.5, which is 2.0R. Third, trades that reach 1R but not 3R, probability 0.5: half banks 1R, the other half exits at breakeven, so 0.5R.

Expectancy of plan B: 0.4 times minus 1, plus 0.1 times 2, plus 0.5 times 0.5. That is minus 0.4 plus 0.2 plus 0.25, which is plus 0.05R.

Ranking: plan A is best at 0.20R, plan B at 0.05R, plan C at minus 0.60R. In this sample, the partial exit sits in the middle, giving up some profit for a smoother ride.

Remember that partials are often justified by behavior, not expectancy. A trader who cannot hold full positions may do better with plan B than with a plan that gets abandoned.

Exercise: recompute with your own hit rates.

Nice calculation. The result for plan B depends a lot on how often trades come back to breakeven after 1R. If many reach 1R and then return to entry, the stop at breakeven saves losses but also cuts the runners.

So the comparison needs a third probability: reach 1R, return to entry, then go to 3R. Without that, plan B looks better or worse than it is.

I went through my own trades with this exact framework a while back, and the third probability mentioned in the reply above was the one that changed the answer for me. About a third of my trades that reached 1R came back to entry before doing anything else. Some of those went on to 3R afterwards, so with the stop at breakeven, plan B was cutting the very runners it was supposed to keep.

When I moved the stop to minus 0.3R instead of breakeven after the partial, plan B improved quite a lot in my numbers. It gave the trade room to retest the entry without being stopped, and the cost on the trades that failed was small because half was already banked.

I agree with the behaviour point too. I tried full exits at 1R for three months, and while the expectancy on paper was better, I kept overriding it when a trade looked strong. The partial plan is the one I actually stick to, and a plan you follow beats a slightly better one you don't.

The calculation for Plan B is careful and correct. The three outcomes, never reaching 1R, reaching 1R and then 3R, and reaching 1R but not 3R, cover all possibilities with probabilities of 0.4, 0.1, and 0.5. The expectancy of plus 0.05R sits between Plan A's plus 0.20R and Plan C's minus 0.60R. So in this sample, the partial exit gives up most of Plan A's edge without gaining enough from the occasional 3R winner. That's a useful result, because many traders assume partial exits are always a good compromise, when the numbers can show otherwise.

The result depends heavily on the hit rates. If 20 percent of trades reached 3R instead of 10 percent, Plan B's expectancy would be 0.4 times minus 1, plus 0.2 times 2, plus 0.4 times 0.5, which is minus 0.4 plus 0.4 plus 0.2, giving plus 0.20R, equal to Plan A. At higher 3R hit rates, Plan B would win. So the best plan depends on how often trades extend. For trend following strategies with frequent large moves, partials can work well. For strategies where most moves stall soon after 1R, taking full profit at 1R is better.

The behavioural point is important. A trader who struggles to hold full positions and often closes them early out of fear might actually get worse results with Plan A or C than the numbers suggest, because they don't execute the plan properly. Plan B, by banking half at 1R, relieves some of that pressure and may lead to better adherence. A plan that's slightly worse on paper but actually followed can beat a better plan that gets abandoned. Testing your own adherence to each plan over a few weeks would show which one you can actually stick with.

Recomputing with your own hit rates, as the exercise suggests, needs MFE data for each trade. Recording how far each trade moved in your favour before exiting, in R, gives the share of trades reaching 1R, 2R, 3R, and so on. With that, you can test any exit plan on paper without changing your live trading. Has anyone here run this comparison with their own data? I'd be interested in whether partials helped or hurt, and how the result changed when different targets were tested, such as 2R instead of 3R for the second half.