Puzzle Answers Aren't Always the Best Move. Why not?

Sometimes when I analyse a puzzle, it turns out that the analysis shows that the opponent in the puzzle didn't play the move the analysis shows as the best, but maybe the second or third best move.

Is this because they are relying on a different engine or depth of analysis?

Thanks in advance for your answer.

Short answer: In puzzles there are no objective "best" moves for the opponent, because they all lose. Long answer: Understanding soundness and motivations in chess puzzles, problems, and studies.

I would say that the puzzle's answer (that is, your moves) are almost always the best moves. But it doesn't always make the best sense for the opponent to play what the engine considers the best moves, because that may not be the best to illustrate the theme and underlying idea of the puzzle. This is particularly true in the situations that arise in the puzzle, when the engine sees that it's losing.

For example, you find a nice mate-in-3. But wait, they have a whole bunch of pieces to throw in front of you, so if they play the "best" moves, you won't even get to play the mate, but just capture a bunch of pieces that get thrown at you, or bat away a bunch of fruitless spite checks.

Or, there's a forced trapping of the enemy queen in 4 moves. But an engine would see that, of course, so instead of your playing moves to trap the queen, you see their queen immediately sac'ing itself at you, taking a minor piece, and take their queen in 1 move instead of 4.

In both cases, instead of your playing out a mate, or playing out a queen hunt, the "best" moves actually make the puzzle easier, and from a human or pedagogical standpoint, incoherent and chaotic.

If you're talking about puzzles here, the initial move you play should always be the best and if it's not, it should be represented.