Accuracy score

I was wondering how does this Accuracy score work. 

A friend got 99.4 and 99.1 %. I know for sure he isn't cheating. But there are other players accusing him of cheating. I was wondering how these scores come to be.

In a short and simple game, some very bad players can play with very high accuracy.  Back when the site was only doing inaccuracy/mistake/blunder, people would post about how they played a "perfect" game because they had none of those.  Nope.  Almost every time the game was short and simple, and a lot of times there were moves that the computer didn't hate, but any good human would reject based on basic principles.  

It means diddly.  But I'm sure it was put in place as a way to gratify some players with some of that good ole "positive feed back" 

Is there anywhere that would explain that getting a lot of best move doesn't mean cheating? Need to some how explain to the other players that just because you get a 99% it doesn't mean they cheated.

https://support.chess.com/article/1135-what-is-accuracy-in-analysis-how-is-it-measured

Thanks

I agree with all these posts. It measures how your move choices line up against the computer's choices. but even fairly modest players can get a high accuracy score for example if the game travels down a well known path, such as Sicilian Dragon Yugoslav - so the first 15 or more moves are all book. Then if one player blunders and the other player spots the blunder and capitalizes on it quickly and accurately (as it may just be basic tactics), then the other player resigns a few moves later - it could be possible for a 1400 rated (say) player to have a 98% accuracy. And there was no cheating going on. 

so you have to be very careful about assuming your opponent is cheating just because the accuracy score seems very high. 

I played a f2f league match this year where my opponent's "centipawn loss" was better than Caruana and Carlson...indeed all past World Champs, up until offering me a draw around move 21 (he had the edge, but I had higher grading...at 1519!).

all because of my simple error early on.

wrote it up in scottish chess newsletter at christmas.

decided relegation ... (maybe...)

i had a 93.1% accuracy as a 400 elo rating player. my opponent had 17.3 accuracy. Im not sure if Im cheating because most of those moves took me about 5 seconds to make best moves and about 6 seconds to blunder.