My thoughts on candidate moves were inspired by reading Kotov, a Grandmaster from mid-century who discussed selection of candidate moves (and many other aspects of interest to chess players). His "Think like a grandmaster" book is, of course, mainly of interest to the chess player but has quite a bit on organizing analytic thought. Anyway, in my prior post I started to get into the idea of "Overlooked candidate moves" (but edited it out). Still, that brought up the memory of Kotov's most famous game versus Averbakh (from the Zurich Candidates Tournament of 1953).
David Bronstein (who later failed to win a world championship when he started daydreaming in a clearly won final game of the match and then made a horrific blunder) wrote:
It is usually thought that the prerequisites of chess creativity are logic, accurate calculation of variation, and technique... There is a fourth component, however, perhaps the most attractive, although it is often forgotten. I have in mind intuition, or, if you like, imagination.
Sometimes positions occur that cannot be evaluated on the basis of general principles...Similarly, a calculation of the variations cannot always be attempted. Suppose that white has six or seven different continuations and that black has five or six replies to any of them.... It is then that intuition, imagination is called into play, which brings to the art of chess its most beautiful combinations and which permits chess players to experience the genuine joy of creation.
Averbakh-Kotov, Candidates Tournament, Zurich, 1953.
30 ... QxP check! (Qxh3 in algebraic notation).
Bronstein again:
It is not true that imaginative games were played only in the time of Morphy, Anderssen, and Tchigorin, and that today everything is based on positional principles and calculations. I am convinced that the games that received beauty prizes in this very tournament were not calculated to the end of all variations. Imagination was and remains one of the foundations of chess creativity...Full game here.
Wikipedia page on Kotov, and page on candidate moves. (The latter says that chess programs have basically abandoned looking for candidate moves, spending their time on more brute force approaches).