Computer Analysis - Stockfish vs. Chess.com

What do you think? Which one is more accurate?

I just analysed a game with both. 

Chess.com says: 11 mistakes, 4 blunders

Stockfish says: 4 mistakes, 2 blunders

So which one is more trust-able?

You haven't said which level of analysis you used on chess.com and you haven't said how long you had stockfish think. I think chess.com uses stockfish 8. Only you know what you used for this test.

Chess.com --> Deep analysis

Stockfish --> Lichess

In that case hard to say since we don't know what hardware it is being run on. I have no experience with analysis on Lichess so anything I say would be suspect. But it will depend on version being run, hardware it is being run on, and how long they take to analyze each move. The are some other variables as well that matter. I also don't know if the criteria for what is a blunder/mistake is the same or not.

But in general chess.com deep analysis is going to be way more than good enough for this kind of analysis unless you are very very good.

And in general, doing engine analysis like this is no substitute for doing your own analysis as well because it has no idea WHY you made the errors you made.

Thats a bad word around these parts.

Yeah I know. I've made my own analysis already, and I was just wondering about said question.

Thank you for your suggestions though

You can look at the depth/ nodes of stockfish and compare which one is better.  I am pretty sure both analysis are quite useless except for blunder check/shallow tactics like 5-6 plys.(in your rating level, it is quite enough though)

For good analysis, you usually need 1 billion nodes per position which is about 3 mins per position on most 4 cores. And a good player usually analyses 3-4 positions per game (10 at most in highly unbalanced/tactical) as they know which move in the game need analysis /which move is turning point of the game.