I just got JBookTrader 9.01 installed and running, I have virtually zero experience on it and basically zero in java... But I was wondering, in order to make a strategy for trading one needs to extended a specific stock(or future,etc) which extends the class strategy. So it seems to me that if you come up with a basic strategy structure that could be applicable to multiple stocks, the only difference being the actual value of the parameter, then you would still need to copy the entire strategy in an new file for each stock, and then say 'public class Sample extends StrategyES' for example.
Wouldn't it be better to decouple the strategy code from the stock itself? Also I thought the Heat map was really cool. It would be nice if it kept the prior saved results and then allowed you to rerun the optimizer with different ranges or parameter steps, in effect 'panning' and 'zooming' the performance picture in parameter space while you find the 'island' of good results. My heat map consisted of just a few short lines, as I had to increase the parameter steps because I ran put of heap memory. But I thought perhaps the map results could be interpolated, like the do in climatology sites, using 'kriging' or some similar method. In any event. For the ES data provided in the distribution, and the initial parameters in the Sample.java file (3200,16,92,-21) I get an off result of $-460 net profit, -0.69 PI after 15 trades. Looking at the heat map it seemed the performance Island was off the upper right corner in the parameter space. Panning the ranges I centered the island (sorry I don't know how else to describe it) as a blob within a window from 3600-4600 in period and in 22.5 to 26.5 in scale, from 98 to 101 in Entry and -6 to 6 in Exit. Choosing the centroid of the blobs as Period 4100, Scale 24, Entry 98, Exit 0, I got improved performance back testing as $2092 net profit, 2.0 PI, after 61 trades This is how one is supposed to choose parameters, right? Thanks for any replies. P.S.: This thing is really interesting, I bet the island moves and changes shape as time goes by, and conditions change. I would like to explore an algorithm that continuously updates the 'centroid' of the island. -- You received this message because you are subscribed to the Google Groups "JBookTrader" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/jbooktrader. For more options, visit https://groups.google.com/groups/opt_out.
