Strictly speaking, chess isn't infinite, but writing about chess might very well be. Whenever I'm working on a new post, I take notes of related ideas, many of which pop up during research. For my recent post, Soltis on Kaufman's Material Imbalances, I noted 'The many references to Spielmann need to be pursued, as does the Kaufman article in the January 2003 Chess Life.' I'll tackle the Kaufman article in this post and return to Spielmann some other time.
'Initial Values of Pieces and Pawns' by IM Larry Kaufman (now a GM) appeared on p.48 of the January 2003 Chess Life. It was a follow-on to his original article 'The Evaluation of Material Imbalances' in the March 1999 issue. I covered that topic in a recent post titled Kaufman's Material Imbalances. For 'Initial Values', Kaufman abandoned his database in favor of engine vs. engine competition, where he altered the initial setup of the pieces, set the engines against each other, and recorded the results. This allowed him to calculate the theoretical values of the various pieces before any moves are made.
The major/minor piece values that Kaufman calculated in 2003 were so close to the 1999 values that it gave him a defacto confirmation of the methodologies used in both exercises. On top of that, he derived a new set of values that I haven't seen anywhere else.
- 1.1 - d/e-Pawn
- 1.0 - c/f-Pawn
- 0.9 - b/g-Pawn
- 0.6 - a/h-Pawn
- 0.4 - Value of a tempo
- 0.2 - Value of first move
- 1.0 - Value of castling
With those values as a reference, Kaufman then analyzed some well known gambits. For example, he determined that the Benko Gambit (1.d4 Nf6 2.c4 c5 3.d5 b5 4.cxb5 a6 5.bxa6 Bxa6) gives Black a better value than at the start of the game before any moves have been made. It might be interesting to try a similar analysis on one or two of the gambits in my Guide to Chess Gambits (Part 1). There I go again -- one post leads to another.
Later: Re '[Kaufman] derived a new set of values that I haven't seen anywhere else', there is some overlap with Lasker's Table of Opening Values. Lasker's values were, however, estimates not based on objective criteria.