I'm trying to build a genetic algorithm, where the strongest logics win and poor logics lose. I'm building a strategy where I'm adding weights to certain logics based on performance. As a results, if logic 1 is outperforming logic2/3/4/5/6 more fund allocation to logic 1. Now suppose, logic 4:6, because it did not get any fund allocation, internally, within the strategy, for a certain time frame, the trades generated from those logics were ignored. Hence, for that time frame, those logics did not generated any trade performance. Hence, through time, bc those logics did not attribute to performance, will theoretically be ignored forever because they have no zero performance.
How can I build a strategy where, even though those logics ignored, the trades generated from those logics are still adding to a "simulated" trade performance, in that way, all of my logics are always contributing to a "simulated" trade performance, and from that trade performance I can attribute funds appropriately.
Any advice or direction would be helpful. Right now the only idea I have is to generate actual trades with a micro quantity for all logics each day and then use those trades to keep track of performance, and then take the best performers for the actual trading within the strategy.
Thanks.
Comment