It’s entirely subjective. Much like in a soccer game when the commentator is commenting on good/bad plays, it’s pretty subjective, but an experienced commentator will have a pretty good idea what a good/bad move looks like.
However, there are also engines that measure centipawn gains/loss. A centipawn is a unit that is defined to be equal to 1/100th of a pawn, so a good (!) move might gain you 50 centipawn units while a bad one (??) might lose you 500. That is what is used for engines which automatically commentate
With that, engines can factor in things like player rating, how many possible good moves were made, time available, etc. to automatically generate puzzles from high level play. Did a GM lose 100 centipawn units on this move? It’s probably due to something really hard to spot, and is a great learning opportunity for novices.
Latest Answers