Go Research on Semeai

• Martin Muller published Race to Capture: Analyzing Semeai in Go.

The article describes the steps of an algorithm to identify semeai. It then classifies semeai in 9 classes (of increasing complexity to solve). It goes on describing semeai evaluation, the safety of stones (does winning the semeai mean that you win the fight?) and determining the game value of winning the semeai. Finally it shows two full board example game positions, that were used to successfully test the algorithm against two world champion go programs.

The article appears to be most useful for go programmers looking for an algorithm to identify and predict fights.

• Robert Jasiek published the book Capturing Races 1 with a complete classification of his "class 1" (mightier than Muller's classes 1 and 2) semeais with two basic groups, a New Semeai Formula and a correction of Lenz's and Muller's earlier semeai formula.

• Thomas Wolf and Teigo Nakamura are other semeai researchers.