matthewvedder / kaggle-rock-paper-scissors

0 stars 0 forks source link

Multi-armed Bandit #3

Open NateSolon opened 3 years ago

NateSolon commented 3 years ago

@matthewvedder

I filled out the stratsdir with some basic strategies. My understanding of MAB is it optimizes over the simple strategies. So for example something like this:

  1. Play random for first 100 moves
  2. Calculate how each of the strategies would have performed against the opponent for the first 100
  3. Start using the one that would have done the best