If that’s the plan, then I guess my next question is how we should go about limiting the strategy space and/or reducing the search quality? (Taking into account things like deception risk.)
I suggested doing this using quantilization.
I suggested doing this using quantilization.