What if you simply enter the market long or short at random, ride your winners and cut your losses?
If random entries can produce a profit then it ought to give practitioners a great deal of confidence in the basic tenets of good trading: running profits and cutting losses. And it is surely a great deal more difficult to over-fit such a system to the data.
In a series of back tests I set out to show that over many thousands of test runs and many hundreds of thousands of trades, random entries coupled with a simple trailing stop could match the average results recorded over the past decade by surviving CTAs.
In the very last series of tests I ran, for the period 2003 to date, the worst test provided the following results:
CAGR 9.29%, MAR 0.28. An MAR of 0.28 is very much in line with the average results for surviving CTAs as a whole over the past decade.
If anyone is interested, they can see the full series of tests run, the code used and the assumptions made at my website.