It's not that rare. In both written and spoken commentaries by experts (super GMs) it is practically standard terminology

"Brilliant move is when you find a good piece sacrifice" - this begs the question, not just for Chess.com but in general, what is a reasonable way to define (algorithmically) a good piece sacrifice? How do you draw the line between a simple recapture, winning back the material by force in a tactically calculable way, and the position being winning in general?

Is there anywhere these rules (regarding discussions/collusion between the games) are rigorously defined or spelt out 'on paper'?

If an algorithm doesn't satisfy you and you need code, then ask on StackOverflow.

@PeterFischer Doesn't koedem's answer satisfy you? Just compare the top move eval to the 2nd best move's eval and set a threshold for what constitutes a 'blunder'.

"Still, the vast majority of those puzzles are terribly easy" - not at all. There are puzzles on these sites that can stump some of the best tactical players in the world (e.g. Nakamura, on Chess.com) for near-on hours, even though they know it's a puzzle. Pose a series of those, midgame, and you'll almost surely win. It's a highly non-trivial question, though where I agree with you is that creating forcing positions and positions which are forcing & difficult are not the same.

Kaufman is of course a tremendous authority on engines and Komodo in particular, but I wonder if this is really true (or if he would even repeat the statement in 2022). Komodo can play blitz well enough to give Knight odds to 2650 GMs (Kaufman himself says so in Chess Board Options). It's hard to imagine any human doing that, no matter how much time they have to think.

@InertialIgnorance Surely a massive understatement. Giving an average GM (or even Carlsen) a lifetime to think would be not much different from giving them an hour or two per move. And already around 2017, the Stockfish team estimated the engine could give Carlsen 1000:1 time-odds.

@Savage47 Not that it's really relevant, since your understanding of the question is nowhere reflected in the question content or title, but I never said "you don't have one rating at a given point in time". I specifically said "your rating vacillates a lot". Now, as for the distinction between the player's rating and their actual playstrength, I also covered this already in my first posts: "if you're playing large numbers of games" (or, precisely, as frequency tends to infinity) the rating will estimate playstrength (insofar as one number represents it) exactly.

Let us continue this discussion in chat.

If your understanding is correct, then the question is 'Why does Elometer / puzzle ability not accurately reflect my rating' or maybe 'What methods best estimate rating without using results/outcomes'. From the question text it's entirely unclear to understand this way, and explicitly contravened by the title (which strongly suggests that the user thinks the 'true' Elo they would get from "an organization" would be poorly estimated by their Chess.com rating).

@lodebari The user has also tried another method that does work, namely, they have a Chess.com rating. I ask now for the third time: Why aren't they satisfied with this estimate? If the answer is that Chess.com rating is distinct from FIDE Elo, then (also repeating for the third time) this should be the explicit and only content of the question as well as the title. Your rephrasing is no more clear than the actual question.

If the question is in fact 'How do I estimate my FIDE Elo from my Chess.com rating', this should be the explicit and only content of the question as well as the title.

@lodebari How is it clear what the question is asking? The problem was never with wording but with content. Issue is simple: User has a Chess.com rating but claims they are not satisfied with this as a measure of their Elo. Why not? What could possibly be better? (not a puzzle website, certainly, let alone one that tests just a handful of puzzles)