Regret Matching

Problem

Convert cumulative regrets into a strategy with regret matching.

Keep only the positive parts and normalize them to sum to 1. If no regret is positive fall back to a uniform distribution, and an empty list returns [].

Hint

Only actions with positive regret get probability under regret matching.

Test cases

case 1

in: [[1,-1]]

expected: [1,0]

case 2

in: [[-1,-2]]

expected: [0.5,0.5]

case 3

in: [[2,1,1]]

expected: [0.5,0.25,0.25]

case 4

in: [[]]

expected: []

Starter codePython

def regret_matching(regrets):
    # Your code here
    pass

4 test cases · run your code against them in the editor