 |
GS
2012-09-02 08:19:46
|
|
Hi,
I am trying to analyze the performance of rules using rules manager and trying to compare it with the one I see using "Simulate/Backtest". This is because, using simulation/backtest, I will not be able to see:
A) total how many signals are generated using a given rule/strategy
B) out of these signals, how many were having highly successful trades (in terms of % return)
C) had I executed all the signals, what would have been the percentage of successful trades
I also observed that the performance of back testing highly depends upon when we are starting the back testing. For example if I am starting the back testing on 1st of a given month and compare it with that when I start on 15th of that month, the performance would very and it is natural and reason is obvious. However, for an individual it is important to know what percentage of his rules are always profitable and overall returns should not have too much of dependency on entry time.
Therefore using rules manager, how do I understand A,B and C? Also, in "simulate/backtest" do we have any way to automatically change the date and time of entry and see how is the performance of a strategy?
For that I tried the following:
Optimize("dd",1,31,1);
SetSimPeriods(2012,1,dd,2012,9,2);
But I could not see any change in the output for different values of "dd". But if I manually change "dd" from 1 to 2, 3, 4, 5 etc., there is difference. Why so?
Within "Optimize" of "simulate/backtest" I could see " buy/sell/long/short timing decalage", what exactly is this function? I could see huge variation in output during its optimization.
Thanks,
|
|
1
