Back testing but 2 sets of results?

jimmy1

Active member
Messages
109
Likes
0
Hi,
I have a system i am testing on tradestation 2000i,
i decided last night i would just use 2 identical charts with exactly the same strategy on it, and guess what? each set of results were different how can this be if the charts were the same and the strategys?? any ideas??? i checked max bars back brokerage commissions and no of days back on the chart? but the p&l and % correct / wrong are different? :devilish: :eek: :eek: :rolleyes:
 
Trader333 said:
Can you post a screenshot of both charts and both Strategy Screens ?


Paul
i can this evening, im at work at the moment, it just wrankles me how this happens i tried it on 3 charts then took the strategy of one chart and then back on and it produced differing results! :eek:
 
Maybe you were optimising one for one criteria and one for another e.g. one for max profit one for min drawdown. Other than that, if all you mention is identical, I have no idea.
 
twalker said:
Maybe you were optimising one for one criteria and one for another e.g. one for max profit one for min drawdown. Other than that, if all you mention is identical, I have no idea.
its weird as both have a set stop and a set profit target, i used to have tradestation 4 and it didnt do this but just speaking to a freind he has the same problem too! :devilish:
 
Have you checked individual signals if the variables are set to the same values? are they the same?
 
twalker said:
Have you checked individual signals if the variables are set to the same values? are they the same?
sorry twalker what do you mean by individual signals?
like make sure the moving average is the same i.e. set at 10 on each strategy etc?
 
I would recommend going through the two sets of results trade by trade (a pain I know). This way you will definitely find the discrepancy. TS can be rather sensistive (especially for some strategies) to some parameters such as MaxBarsBack and the number of days of data loaded in for example.
 
Jimmy1,

My experience is that the most common cause for your problem, provided dates etc are the same, is using different strategy testing resolutions (under Properties->Backtesting Settings). Check that they are the same.

Paul
 
fowkesp said:
Jimmy1,

My experience is that the most common cause for your problem, provided dates etc are the same, is using different strategy testing resolutions (under Properties->Backtesting Settings). Check that they are the same.

Paul
hi and thanks both for your help much appreciated i will look into :D
 
Top