The concept is not that complicated but generally many don't consider it. That included myself. I was using something similar but not same as what is covered in this post though - a topic for a future post.
First we need a strategy before we can improve upon its performance. Any one of the numerous studies posted on this blog will fit the bill. But it is more fun doing a new strategy. So below is a strategy with rules to capitalize on a old concept - Turn around Tuesday
Turn around Tuesday
The idea is simple. Tuesday's are known for markets to reverse their direction. So our setup builds on that with a small twist.
Buy Rules:
- Today is Monday.
- Today is a down day i.e., Close < Open.
- If above two conditions are met, buy @ open the following day (i.e., Tuesday Open)
- Sell the following day i.e., on Wednesday open.
Usual caveats - results are frictionless i.e., no slippage and no commissions. The equity curve is a closed equity curve. Also any information posted on this blog is not a recommendation.
Results:
Following annotated images provides the performance stats and my thoughts on the system performance.
Now the question is can we improve the results of the above system? Who doesn't like getting rid of some of the losing trades. A penny saved is penny gained and savings from avoided losses adds to bottom line.
One default approach taken by many is to try add one or more market filters to improve performance i.e., to reduce number of losing trades and increase number of winning trades. But another equally interesting option (IMO) is managing the equity curve of the system. In other words, we let strategy footprints also provide feedback along with market footprints to guide us in making trade decisions.
Equity Curve Management
The idea is fairly simple. Add a short term, mid term and long term moving average to the equity curve generated by the strategy. Just like markets, strategies also goes through sideways periods. So using multiple MA's provides a better picture.
Rules:
- Create a 10 trade moving average of the current equity curve.
- If current equity curve is above its 10 trade moving average, and a new signal is generated (that meets all strategy rules mentioned above) then take the trade.
- If current equity curve is below its 10 trade moving average, then stay in cash and paper trade an new signals and update the strategy equity curve.
- If you prefer longer moving averages then instead of 10 trade average, do the above with bigger average like 30 MA or 50 MA.
- There is nothing special about 10, 30 or 50 MA's. I just picked them randomly. Also it is not necessary one has to use only moving averages. There are lot of other technical indicators one can use and may be depending on reader interest we cover some in future posts.
- In this post, the equity curve management is based on concept - follow the trend.
Some people prefer to double down on the trade size when the equity curve comes below its moving average. In other words, applying mean-reversion concept to the equity curve.
Following annotated images provides the performance statistics for all three scenarios i.e., 10 MA, 30 MA and 50 MA based equity curve management along with my thoughts on the system performance.
Wish you all good health and good trading!
4 comments:
Hello, just curious, why do you say "The concept is not that complicated but generally many don't consider it. That included myself."
because according to your statistics, it seems a promising idea.
Thank you.
Hi Anonymous - Good question. The reason I felt that many don't consider it is because we don't see often people talking/writing about using Equity Curve management. In my case, I heard about it but didn't pay much attention to it before. After the tests, I too felt it is a promising idea and plan to include in one of my systems this year.
Hello, I got to your article just now and I tried to implement something similar to my strategy.
I believe you made a mistake of foresight - using knowledge of current trade to compute outcome of it (whether to take it or not).
I made the same error when I tried to compute it in Excel and I got perfect equity curve for ECM.10 only to later find my mistake.
For example the very first loss taken by original strategy and avoided by ECM.10 wouldn't really be avoided because equity wouldn't be below 10 MA.
It would be below 10 MA only if you'll count in also the losing trade that was avoided.
I hope you understand what I mean. This method is very interesting, in my tests it can reduce drawdowns, but it doesn't increase net profit.
If I'm wrong would you be so kind to show me your data? What tool did you use to run these tests?
Well spotted Mark and good valid points (backed up by our research).
Post a Comment