12,660 trend-change arrows replayed: a fixed target lost, break-even plus a trailing stop did not
I replayed 12,660 trend-change arrows (a Supertrend-type line, ATR 10 x 3) on the history of USDJPY, EURUSD, GBPUSD and XAUUSD, and compared two ways of leaving the trade. The result surprised me enough to change how my indicator works, so here are the numbers.
The test
- Entry: the close of the bar that confirms the change of direction.
- First stop: the low of that bar for a buy, the high for a sell (at least 0.5 x ATR away).
- The order of touches inside a bar was decided on M5 data. The spread was subtracted from every trade.
- R = the distance from the entry to the first stop. A result of +1R means the trade made as much as it risked.
Two exit rules
- Fixed target: take profit at 2R, or the stop, or the next arrow.
- Break-even and trail: when a bar closes at +1R or better, move the stop to the entry price. After that the trend line is the stop. No target.
Results, 4 symbols pooled
| Timeframe | Exit rule | Arrows | Average | Profit factor |
|---|---|---|---|---|
| M15 | Take profit at 2R | 9,740 | -0.109R | 0.84 |
| M15 | Stop to entry at +1R, then trail | 9,740 | +0.016R | 1.07 |
| H1 | Take profit at 2R | 2,332 | -0.057R | 0.91 |
| H1 | Stop to entry at +1R, then trail | 2,332 | +0.068R | 1.44 |
| H4 | Take profit at 2R | 588 | -0.035R | 0.94 |
| H4 | Stop to entry at +1R, then trail | 588 | +0.267R | 2.87 |
What the numbers say
- The fixed target won more often, about one arrow in three, and its average was still negative on every timeframe.
- With the break-even move and the trail, only about 8% of the arrows ended in profit. Most of the rest ended as a small loss or close to zero. A few large moves carried the whole result.
- On H1 the average was almost the same in the first and in the second half of the history (+0.069R and +0.067R). M15 was weak: its first half was about zero.
- Not every chart was positive. Some symbols and timeframes had a negative average with both rules.
What I would not conclude
This is a replay of past prices. It says nothing certain about the future, and real fills, spreads and swaps differ between brokers. A rule that wins 8% of the time is also hard to follow: long rows of small losses are normal for it. The useful part for me was the comparison: on these arrows, cutting the winners at a fixed target cost more than it protected.
How I use it
My indicator RiskTrend Planner draws this plan on every arrow (entry, stop, the level where the stop moves to the entry) and replays the past arrows of the chart you are looking at with the same rule, losses included. When the past average of that chart is not positive, it shows a suggested risk of 0%. It is a planning tool: it does not open trades and it makes no promise of profit.
