Retrograde Analysis and Two Computerizable Definitions of the Quality of Chess Games

Abstract

This paper deals with a twofold problem: first, that of the realistic content of chess problems. Second, retrograde analysis is shown to have fundamental questions of its own: notably, given two configurations A and B, is there a legal chess game leading from A to B? No program seems to be available to decide the matter. As a research question, it is proposed to implement if possible a definable quality, modulo X, where X is a machine. In turn, this gives rise to the question: ‘Is there a fair procedure to evaluate the strength of a player, P, from the moves played, disregarding the rating of P’s opponents?’.

The paper intentionally proposes research questions without furnishing specific answers.