Evaluating the Results
Previous  Top  Next

In our run, looking at the Search Space indicator, we can see that the search space was covered fairly evenly, and probably with enough penetration to discover the major features of the fitness landscape. Wider purple rectangles indicate areas in which the search was concentrated.

sample1_sspacescalarmap
The search space indicator shown here indicates fairly even search space coverage.

With the ISpace indicators, we can see that favorable values for Len1 cluster around 42, and values for Len2 cluster around 8. To see an indicator clearly, click it as shown here:

ispace
ISpace indicators display a heat map of the evolutionary process.

Here we see the input space detail canvas for the Len1 input.

sample1_inp1
The Input Space indicator for the Len1 input indicates a primary clustering of favorable inputs around value 43.

Here we see the input space detail canvas for the Len2 input.

sample1_inp2
The Input Space indicator for the Len2 input indicates a solitary clustering of favorable inputs around value 8.

Returning to TradeStation and plugging Len1=43 and Len2=8 into the inputs for the strategy (while setting OMX_IterationNum, OMX_Generation and OMX_Individual to 0) yields the following equity curve:

sample1_equity_curve

Several trials using nearby values for Len1 and Len2 also yield similar equity curves. This means that Optimax has identified a stable area in the fitness landscape with smooth approaches to a maximum.

Note:
Due to the nature of genetic searches, it is possible that your results will not be exactly the same as ours. You should, however, be able to obtain comparable results. Running the optimization several times may yield different results due to the randomness inherent in the search process. In any event, you should be able to obtain satisfactory results when Optimax is used correctly.