Let the AlphaGo Teaching Tool help you find new and creative ways of playing Go.

Read MoreRead Less

This tool provides analysis of 6,000 of the most popular opening sequences from the recent history of Go, using data from 231,000 human games and 75 games AlphaGo played against human players.

Explore the board and learn how AlphaGo's moves compare to those of professional and amateur players.

Overlay: None AlphaGo Value Rank Occurrences

Key

  • Moves AlphaGo would play
  • Moves human players would play
  • 46.8AlphaGo's evaluation of the move for black's winrate
  • Last move played

How to Use The Tool

Use the coloured circles or the controls to navigate the board, explore different opening sequences, and find out AlphaGo's estimate of black winning for each move played.

The number in each circle is that move's % winrate from black's perspective, based on AlphaGo's predictions. When it is black's move, values closer to 100 are considered better, and when it is white's move, values closer to 0 are considered better. 50 means the game is considered equal.

Understanding Alphago's predicted winrate

AlphaGo’s preferred move does not always have the highest value. This is because each move’s winning probability was computed by running an independent search of 10 million simulations from that position. AlphaGo has some randomness in this search, meaning it might pick different but similar value moves if we run the search again. For more information on how AlphaGo works, read the original paper published in Nature.