Genetic "algo"

[tommiehansen]

absolutly right.
i was thinking of genetic-js to manage the population mutations and cross overs. Almost the whole complexity ist in genetic.js file without external dependencies. https://github.com/subprotocol/genetic-j...genetic.js

Each Run:
{
for each entity it generates, will start a backtest in gekko and get "sharpe ratio" as fitness factor back.

genetic.fitness = function(entity){ call Gekko(entity);  GetSharpeRatio();  return sharpe ratio; }
https://github.com/subprotocol/genetic-j...itter.html

after the run, using the result on some random Dates. (to test if the result not just works on  sample Dataset)
}
having many runs and comparing the Results.

similar Algorithm from Gekkowarez: https://github.com/gekkowarez/gekkoga/bl...e/index.js (we shoul find here the communication part with gekko)

does it make sense?

No, because you skipped anything that has to do with saving/ouputting/formatting/viewing/user input.
Basically the entire way of actually getting something out of that script.


If you like it why don't you create something out of it?

It seems that you underestimate what one has to do in order to get stuff running.
I want it to be as simple as possible, then one can start thinking about adding complexity.