In SQX another useful tool is the genetic building process. This process can shorten the building time immensely and with the right settings, can do so without inducing significant curve fitting.
Sadly, currently we cannot combine the two tools and get 100% perfect functionality because sometimes crossover winds up not respecting the templates indicator groups or the indicator or templates variable ranges. (https://roadmap.strategyquant.com/tasks/sq4_7824 ; https://roadmap.strategyquant.com/tasks/sq4_010067 ; https://roadmap.strategyquant.com/tasks/sq4_8856 ; https://roadmap.strategyquant.com/tasks/sq4_9768 ;)
This request is just a general request for the SQX team to solve this issue in any way they see fit.
Crossover is allowed to bring in variables from one parent and insert them ANYWHERE in to the child. So for instance, the level for the RSI trigger could become the deviation for the Bollinger Band of the child which is nonsense and is causing many crossovers to fail wasting CPU and reducing efficiency of the migration/island system (even if not using a template). Please consider making some changes to improve this.
So perhaps one of these or combination:
-completely redesign the crossover system
-introducing an optional check of all crossovers to make sure they are in bounds of the template. In this way it would only slow down the building if we turn the optional check on.
-https://roadmap.strategyquant.com/tasks/sq4_8897 (Make children that have exit rules from one parent and entry rules from the other.)
-https://roadmap.strategyquant.com/tasks/sq4_9808 (Restrict crossed over blocks to be exact same use for parent to child)
-crossover only blocks that are not restricted by the template except for the purpose of same use.
-check all strategies after they are produced to make sure they fit within the template, if not then delete it.
-allow the community to create their own genetic evo algorithms https://roadmap.strategyquant.com/tasks/sq4_7689
Please post your ideas to solve this. It is a pretty big shortcoming that we cannot get full functioned templates in use with genetic building.