Quote:
Originally Posted by robbi
It was 90%.(offline testing, historydata of good quality is on my harddrive)
You are writing in the description of this EA: "...instead you need to optimize it to the market to set the most fitting parameters"
I did that not, because I wanted to know, to what the parameters of your testing-period (11/19/2006–11/19/2007) would result in the time after your testing-period.
Nobody knows the market-condition in the near future, so MyRandom is a very funny EA, but thats in an other thread, so I stop here.
|
Yeah, Artificial Intelligence needs to be optimized quite often. Like, if you are going to test it on some one year period it's better to optimize it on a previous year before doing a test. If you are forward-testing then there is no problem in doing constant re-optimization every weekend. The more up-to-date the parameters are the better will be the results, in theory
.