Managing Trend-Following Drawdowns and Testing Overlays
Summary
The document considers ways to make a trend-following system’s equity curve less volatile after observing substantial drawdowns. Suggestions include studying whether returns show persistence or reversal, using a secondary trend or equity-based stop to adjust exposure, testing entry filters, and exploring supervised learning or control-chart ideas to identify unfavorable periods. Reducing position size is also presented as a practical way to limit risk.
These overlays may make losses more tolerable, but they add model risk and may fail to improve results. Parameter optimization, including changes to exits, can produce misleading backtest gains through overfitting. The discussion recommends checking proposed improvements on separate in-sample and out-of-sample data. It also notes that trend-following strategies can have larger drawdowns than mean-reversion approaches, without providing comparative data or a tested solution.
Key ideas
- Analyze return sequences to assess whether performance conditions persist or reverse.
- Test exposure overlays, equity-based stops, entry filters, and learning methods cautiously.
- Reducing position size is a direct way to lower the system’s risk.
- Parameter tuning can overfit historical results, so evaluate changes out of sample.
- An overlay can compound model risk and may not improve performance.
Tags
Full text
# System Development / Optimization # System Development / Optimization I have been testing a trend following strategy. The results shows massive drawdowns which makes the equity curve very unstable. I just wanted to know what are some ways in which I can reduce the volatility (increase smoothness) of the equity curve? Better exits? Better entries? ## Answer by Joshua Chance (score 6, accepted) https://quant.stackexchange.com/a/486 Another possibility is to analyze the equity curve itself so as to go live with the system when good performance is expected and to either reduce risk or just paper trade when performance is expected to be negative. Are a series of positive returns followed by negative returns (i.e. is there mean reversion)? Does a trend-following "meta-system" and/or a trailing stop loss (on the total account equity) reduce risk or at least make it more tolerable? A couple other ideas might be to try supervised learning for the drawdown periods or incorporate concepts from control charts. There is some danger that you might just end up multiplying your model risk by overlaying a meta-strategy on your original system. Statistically significant changes may be very hard to come by and in the end you might just have to trade smaller size. Good luck. ## Answer by Zarbouzou (score 4) https://quant.stackexchange.com/a/484 Be careful when you optimize the exit parameters (and any other parameter) as you could get better results in backtest that will only be due to over fitting. IF you haven't done that yetn use In and Out sample to verify your improvements. After that you can try to build entry filters. In my experience trend following usually have bigger drawdowns than mean reverting.
Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.