Azmel,
I did some preliminary backtests yesterday. While I did not get the modeling quality I would have hoped for I did find the following:
TP & SL can definitely be optimized but are not too far off.
The main problem hindering profitability is the Stoch upper and lower targets. They are at 80 and 20 now but altering them can make this a very profitable EA--versus a losing one. Unfortunately, a minor variation in targets can alter the results considerably. So I quit testing with the thought that I was overoptimizing and probably couldn't replicate forward what had occured backward.
However my message is, that even though changing the stoch targets runs counter to the manual method, I couldn't find the EA to be profiable without changing stochs. So for others to help, you may want to offer stoch upper and lower as external (testable) parameters.
saintmo
Quote:
Originally Posted by azmel
Thank you for providing the back-test results, janad.
I propose that we optimise this EA. I know that OzFx state StopLoss and TakeProfit of 50, but these figures are not official. I have attached the following v1.02 EA in beta mode which you can change the SL and TP values.
I need volunteers with good backtest data (90% modelling quality or better) to optimise the EA and find the optimum SL and TP value for all pairs.
I know this is a big job as there are 28 possible pairs, so please let everyone know which pair you are optimising so that it won't be duplicated by others. Once you have found the optimum setting, please report back.
Thank you all.
Azmel.
|