How important is accuracy? ( in post game analysis)

This might initially sounds like a stupid question and maybe it is  But how important is accuracy and what is a good number? I get lots of games in 90+%, is that good. In most aspects of life 90% is good, but in chess that could be garbage? I'm a new player, hoping to improve. I do analyze most of my games and I was wondering about that accuracy number.

Thanks, and if this was a dumb question, I apologize

Part of it depends on your opponent. The better they play, the more difficult it will be to get a high accuracy percent.

Part of it depends on the opening. Some openings lead to stale and straight forward positions while other openings lead to such crazy positions that it will cause even very good players to make a lot of mistakes.

Part of it depends on if the positions lend themselves to super quick analysis (which is what chess.com is doing). For example if you play a move that is mate in 6 but the engine sees a way for you to play a mate in 5, then it can say you were inaccurate even though this sort of inaccuracy doesn't matter. Also in some endgames and some attacks the correct moves involve very long term thinking, so the quick analysis can mark your good moves as inaccurate or your inaccurate moves as good.

---

So the number is not really that important, but it is a fun feature, and getting a high percentage is always nice even if it doesn't actually mean you played like a GM most of the time

Mostly I take it as meaning I played a lot of reasonable moves without many (or any) big blunders. Were they the best moves? Almost certainly not, but consistently playing very reasonable moves will put a lot of pressure on human opponents, so it's nice.

those were essentially my thoughts as well, but you explained it much better than I would have  Thanks for the comment.

@llama44 - thank you for your explanation. I certainly hope a lot of people here read it or else they might submit reports to the staff based on these high figures. Personally, I'd prefer it if the accuracy percentage was calculated by simply deducting the inaccuracies, mistakes, and blunders from the total number of moves and working out the accuracy based solely on good moves.

I've been told it's easiest to find the correct moves in tactical positions than positional positions. My endgame play is not good enough for me to offer any idea how challenging it is to play it correctly. During a recent video, GM Delorme mentioned in practical play the game would most likely end in a draw but with the use of tablebases, white would win in 67 moves (Queen and pawn versus queen). Good luck getting a high accuracy rating in that position.

In a recent blitz game lasting 27 moves, I scored 98.1% accuracy. The game report showed me as making 2 inaccuracies and 2 mistakes. I will consider my accuracy at 85.1% (23/27 good moves). This includes a five move mating sequence any reasonable player could spot.