By misstake i found the following. I had loaded strategies twice in SQx and got the _(1) addon and then save them. Now i had build a portfolio using the same strategies but wanted to improve it by adding more so i Merge the portfolio to single strategy and ran it again. Now even with the correlation setting of 0.3 i had 10 new strategies added that already was in the merged portfolio without the _(1) which have a correlation of 1. I found out first trading them on real account since i had them 4 of them taking the same trade !. You owe me many Beers now :)
Yes that has been my thought aswell. Maybe i had to many strategies in the databank and it flipped out or maybe the merged portfolio had to many trades or maybe it was beceause i still had the merged singel strategies still in the databank which i did. Donno but it happend and my portfolio have strategies wich are identical because of this. At two previous occasion during last 3-4 years i have complained about correlation issues which destroys portfolios perfromance when trading live and you have to manually correct remove rezise portfolios during live trading because you trusted that QA result was correct. Both times it was first said that it was already working correct but then later it was fixed and the last bug find it was a great improvement. However this is difficult and correlation is changing by market condition but this might be bug need to be looked into since in current version of QA it is difficult to impossible to work with many strategies and you need to split them up i smaller portfolios and work with those.
imagine that you are computing monthly correl - its obvious that monthly profit/loss will be different for only 1 single strategy and for the whole portfolio as it is merged to a single strategy
solution is only one - rewrite the correlation compution for merged portfolio so the correlation will be computed for single strategies in it
merged portfolio behave as a single strategy