Reminder for myself to finish automation for validating performance metrics. We currently lack good tools to test if the replacement math is working better than vanilla. I can currently run unit tests locally but results are very random due to other uncontrolled factors. To remove this issue we need to boot up the mod in a docker image and run the alg under a series of setups. To include in this we need to test each major JVM flavor, several forge versions, different OS if we can do it for free, and some of the top downloaded AI/Entity mods.

Additionally we can use this to bug hunt for mod conflicts.