Hey, Game theory geeks.
When using regret minimization to find what Player 1's strategy for next iteration is, should we add negative regrets to the Sum regrets, or should we only work with possitive regret?
I guess both would work but the latter would take more iterations to get to the same precision as the former?
