Workshop7 does not work in reality.
So the question is, wouldn't it make sense instead of train on real data, do the train on the DETREND = SHUFFLE a 1000 times and then filter the PATTERNS afterwards on their succes on the real price curve?
That could significantly speed up development process, cause a humen does not need to be involved.
Since Zorro does not support it to verify my thought, I would need to somehow store the generated paaterns, e.g. copy to a new file.
Now if running the collected PATTERNS vs. the real price curve, how would I identify the succesful patterns?
It would definitly not work to test them one by one.
As far I see, the log does not offer any hint about the patterns used.
I tried Verbose = 7+DIAG;
First test show, there might be some validity to the thought.
@Zorro Team, is it possible to make the information of the optIF available per Pattern?
That would be probably something like a BruteForce PATTERN search.
Last edited by danatrader; 05/30/20 09:12.