Why are people frequently playing 400 points above their rating?

So I don't always play well and can blunder or miss winning positions.  But quite often I do play quite well but get beaten in the mid-game, only to find from analysis that my opponent was playing 400 points above their rating.  Anyone know what's going on?  Is it engine cheating, or cheating from secondary accounts from players who want to practice traps, or is it just natural variation, or are the ratings wrong? 

When I run a game review, the engine tells me the equivalent rating that I played at and that my opponent played at in that game. I save my reviews for tough losses, and I often find my opponent might have played at 1800 according to game review, despite being rated on chess.com at 1400 (rapid) or 1600 despite being rated 1200 (blitz). Maybe its a selection bias or selective memory? But it seems odd to me. Often I have to play hundreds above my rating (according to the game review rating) to win.

You don't really know he was playing 400 points above his rating. Probably just some silly assessment by the engine. My 5 min blitz rating is 1530 at the moment, because I'm slowing down a bit, mainly. I probably play games at 2200 when I'm playing well and sometimes my rating dips to 1400, so I'm perfectly capable of playing at 800 points above my rating.

Your rating in a game is hardly ever going to match your rating at the time. My game reviews have recently told me I played like a 2000 and then in another game like an 850.

Oh I'm a non-paying member so I'm limited to one or two game reviews per day but I never use them at all unless I want a cheap laugh. We know very well that sometimes we play better or worse than at other times.

I think the ratings are just an average score or even oblique and not a true assessment.

For instance the ratings do not take into account if you has had a glass or two of wine )))

The game review estimated rating is just an estimate and not accurate. It doesn't mean they were actually playing at that rating.

The calculation takes the ratings of both opponents and the accuracies into account. The exact same game played by different ratings will have different estimates.

If you are talking about the estimation given by chess.com, don't put too much stock in it. Both accuracy and rating estimation depends on the rating of the player. The same game from you and some 500 rated player will not be rated the same (your scores will be lower than those of lower rated player). Here are some obvious examples (these were unrated games against lower rated players) where rating estimation is ridiculous:

9 moves game, the opponent drops a piece and gets 77.5 accuracy and 1 400 estimated rating (while being below 900). People around 1 400 rating don't usually play like this.

8 moves game, the opponent blunders a pawn, then loses a pawn and an exchange, 61.9 accuracy and 1 000 estimated rating, while being around 500.

14 moves game, the opponent drops a piece for a pawn and gets 74.1 accuracy with 1350 estimated rating, while being just below 1 000 rating.
And so on... There are certainly examples of longer games as well, but these are just most clear cut examples.
Sadly, I can't find a favorite game of mine. In that one, the game lasted around 9 moves, there were 3 blunders and the opponent gets just below 70 accuracy and like 1 200 or so estimated rating.

Basically, you need to actually look at the games in detail to judge them yourself, accuracy and estimated rating are pretty unreliable and subjective because they are somehow dependent on the rating of the players.

Edit: Oh yeah, and don't get me started on those games where you exchange most of the pieces, get to the relatively easily drawn endgame and mostly play one of the many safe moves. You will get around or over 90 accuracy in many cases and the estimation is 2 000 or more. On my level at least, in those type of games, report tends to overestimate player's strength a lot.
Here is a recent one:

Exchanges happen, the opponent has a better position after I misplay it a bit, and eventually I am a pawn down. Then I somehow get an opposite color bishop endgame with a pawn down and salvage a draw. this is a game where you basically put the pawns on the color of your bishop and with it and your king are often able to defend everything, so it is not that difficult not to blunder, especially when you have a lot of time. Sure, I am pleased that I did that, after not playing well in the opening, but let's see the estimation.

The result, we both get 96 accuracy and the estimation is 2450 for both. Titled Tuesday, here I come... or perhaps game report exaggerated "a bit"...

Here's my last two games. 1400 in a win and 650 in loss. I played well, and thus the 1400. I blundered like a fool and therefore 650 LOL

I can't get back up to 700 because I'm always playing people that are rated 1400+ in game review

Says the guy who made a post that said "I'm downloading Stockfish." 🤷🏻‍♂️

You're so stupid you have to explain

I'm not from the canary islands but you being racist is quite funny, also, you avoiding answering is pretty funny.

Blud probably has the one inch millimeter defeater

You are being racist because you hated on me because of the place you thought I was from, and then you resorted to disrespecting my mother because you are coping for the loss of your mother after she left