What is the difference between blunders, mistakes and inaccuracies in chess mean?

@CM Sarg0n by your math terminology KID is a mistake (stockfish is indicating +1) Benoni also nad many other opening lines, but they are good. Mathematicians who used to count everything failed in chess. It was much more easier to soviet GMs to sac a rook for bishop, because they didn't treated everything by math on the chessboard.

You can actually check back #5 from what I said.

@StingerPuzzles thats because the position is still completely winning even if u sac ur queen but if u were to sac ur queen in an equal position it would be a blunder it all depends on the position

I think this is the relevant source file: github.com/ornicar/lila/blob/master/modules/analyse/src/main/Advice.scala. I could try to explain it step by step but I'm a bit too lazy for that.

@StingerPuzzles The answer to post #24 is in line 125 of the above source file. When you go from mate to a position between +7 and +10, that is considered a mistake (this is hardcoded).

This is a verbal description from old books. You have to extract the meaning from the context. If you believe you can define it by engine evaluations you're lost to critical thinking.

With 16 you should have got that. I'm afraid of a humanity where such a question is discussed with so much answers instead simply stating, with some more words for sake of politeness: Think yourself!

Thanks, #34. It’s dirt-cheap mathematics here. What is it what you don’t understand?

@PHlinka

losing forced checkmate sequence is not always a blunder for the computer, it depends on the position:

look at this game - i missed mate and hung a queen, but computer analysis said i made no mistakes:

Sarg0n Blunders, mistakes and inaccuracies are not explained by mathematics. What do you don't understand here? I don't have time to arguing with you, so stop here and stick to your math :)

A position can only be won or it is a draw. An error changes the result (to the bad of course). Win to draw or loss, draw to loss.

So what do you mean by inaccuracy? Does it change the result? A mistake? A blunder?

We‘re talking lichess here. And here it is just mathematics.

We found this in the source code:

private def cpWinningChances(cp: Double): Double = 2 / (1 + Math.exp(-0.004 * cp)) - 1

private val winningChanceJudgements = List(
.3 -> Advice.Judgement.Blunder,
.2 -> Advice.Judgement.Mistake,
.1 -> Advice.Judgement.Inaccuracy
)
cp = centi-pawn

When you plot the formula (www.wolframalpha.com/input/?i=2+%2F+%281+%2B+exp%28-0.4+*+x%29%29+-+1+from+0+to+4) you see that a blunder roughly starts at a drop of 2, a mistake between 1 and 2 and an inaccuracy above 0.5, just as indicated in post #30