Steve2 Elite    Posts: 750 Joined: 10/11/2012 Location: Annapolis, MD  User Profile | To continue the mucking... You can verify what Mark is saying by looking at the ECA Dump data for a given PW run. It will show you the build up of strategies and the EF values for every combination that was considered. What I noticed when looking at this is that the number of strategies that produce the highest EF value for a given period is not necessarily the number of strategies that you specified in the PW Analyzer Settings page. For example if you had set Strategies to 10, it might be the case that for a given period a combination of 7 strategies produced the highest EF value and by adding in 3 more strategies, PW produced a lower EF value. This suggests a way to optimize the selection of the value for number of Strategies by performing a PW run that selects n-1 of n strategies (where n is the number of strategies that are enabled for the run). Then examine the ECA Dump output and identify, for each period, the number of strategies in the combination that produced the highest EF value. You then average the numbers across all periods and if the standard deviation is not too high pick this average number as the PW Analyzer Settings for number of Strategies. This saves you from having to do a number of PW runs across different numbers of strategies. To do this though, you need to dump the ECA Dump data into excel. Steve [Edited by Steve2 on 1/9/2015 12:51 PM] | |
