Posted: Wed Mar 28, 2012 3:31 am
After accidentally stumbling over this interesting thread here I was reminded of my own random entry/exit-experiments. After all, I'm pretty sure sluggo made a mistake in his tests.
As he found out (correctly) the trading-rule-set has to overcome a certain annual growth rate just to cover transaction costs. However transaction costs always directly depend on the number of trades. So his reference test would - in my eyes - only be a valid reference if the number of trades of the following tests would at least roughly match.
In my tests I tried to create 'comparable' references for every test I made by creating random entry-random exit-tests with a comparable amount of trades. (It was a long time ago, now I would rather use a simple formula I guess.)
Please correct me if I'm wrong (or if this is so obvious that no-one found it worth mentioning...).
As he found out (correctly) the trading-rule-set has to overcome a certain annual growth rate just to cover transaction costs. However transaction costs always directly depend on the number of trades. So his reference test would - in my eyes - only be a valid reference if the number of trades of the following tests would at least roughly match.
In my tests I tried to create 'comparable' references for every test I made by creating random entry-random exit-tests with a comparable amount of trades. (It was a long time ago, now I would rather use a simple formula I guess.)
Please correct me if I'm wrong (or if this is so obvious that no-one found it worth mentioning...).