Contrast and Condensation in Analysis of Chess Games

Abstract

We created 4 sequences of chess moves intended (and verified) as sufficiently good (2 sequences), in play quality, or bad (2 sequences) to induce contrast. In experiment 1, 24 experienced chess-players (USCF Elo > 1300) watched these sequences and rated them with regards to overall quality (−100 to +100) and estimated Elo ratings, a proxy for play quality, of the players involved. In experiment 2, a different group of 24 experienced chess-players rated the sequences of chess games by “How much better did the winner play than the loser?” on a 1 to 7 scale. Results revealed negative contrast (experiment 1) and no evidence of condensation (experiments 1 and 2) as well as the potential that one’s own actual Elo may have anchored the ratings given to one set of stimuli.