CODEWITHSUNDEEP
artificial-intelligenceminimaxalpha-beta-pruninggame-theory
him
him

Reputation: 489

α and β in αβ-pruning

I understood the algorithm for αβ-pruning in minimax calculations. However, when I was trying to have an intuitive understanding of why the algorithm actually works, I was not able to come up with a satisfactory meaning for the value of α and β values at a node.

So, I just wanted to know the significance of these values - which might help to get a clear intuitive understanding of the algorithm.

Upvotes: 1

Views: 81

Answers (1)

Xteven
Xteven

Reputation: 441

TL;DR: The α and β values you have when searching a game tree are the already guaranteed worst scores (for that particular player) that a player can achieve.

By convention, the α value is the best 100% guaranteed value of the maximizing player (initially at -infinity) and the β value is the best 100% guaranteed value of the minimizing player (initially at infinity). These values come in use when searching the game tree because if you ever find a value of a branch of the tree to be worse for the particular player than their already guaranteed score, you do not need to search further as any following moves won't be played.

Since I am new to stackoverflow, please tell me how I can improve my answer.

Upvotes: 1

Related Questions