OK I'M DONE

I'm done with Chess.com puzzles. It's just outrageous. The rating system makes no sense, I've been 4 months stuck around 1700 rating bc whenever I climb up, a bunch of 1100-1400 rated, harder than their rating puzzles will show up to crush me down again. I am not one of those who usually complain and leave it at the first time, but it has reached a point where it is discouraging and tiring. The straw that has broken the camel's back is this puzzle. Scored 1400 (what a surprise). Literally the module gives a draw at the start and end of the puzzle (there is no best move, it should be a draw by repetition). The move that 'solves' the puzzle involves you predicting that the opponent will break the 3 move draw by making a move that disadvantages him (and put him in Mate in 1), so you just keep moving your rook in hopes that this is what will happen. It's moronic.

I'm fed up, congratulations Chess.com you have discouraged one more amateur to enter the world of chess. You should be ashamed of yourselves.

It doesn't matter if the opponent breaks the repetition or not. The whole point of the puzzle is to see the only option is to keep checking. After the deviation white has to still step back to g3 and the position will repeat.

https://www.chess.com/puzzles/problem/534118

Don't let it get you down.. There are crappy puzzles here and there. You'll get a surprisingly easy one that balances this one out. I'm stuck in the 2500-2800 range and can't crawl out either.

Some puzzles are over-rated. This one is a Queen mate-in-one for 1900:

https://www.chess.com/puzzles/problem/2334098

Boohoo

I have just tried the puzzle. Literally nothing that OriolAlegria said is true. It's just a straightforward draw puzzle, and none of those weird things happen.

Yeah you're the sharpest stick in the shed, thanks for the memo. I just said that I think that puzzle should not be rated 1400. 5 out of 6 commenters got it right.

No, you didn't say that. You said

"the move that 'solves' the puzzle involves you predicting that the opponent will break the 3 move draw by making a move that disadvantages him (and put him in Mate in 1"

That doesn't happens in the puzzle.

Yeah you got that wrong. I said 'solves' with quotation marks because I know that is not the solution, however, when a 1400s faces the problem she's not thinking that repetition is the solution but rather some way to mate in 3, which is misleading. Not for a 1900s, but for a 1400s.

The problem with your explanation is that anyone can see what you wrote originally.

Puzzles are given within a relatively narrow range of a player's rating, so even if there is occasional cheating on puzzles, it's not going to have a massive effect on the ratings.

Also, I believe the site locks puzzle ratings at some point, so they don't change again unless manually modified.

No, he's right the rating system is bizarre. The puzzles do not seem to change rating based on success rate on player attempts relative to the players' own puzzle rating. Which is to say the rating system for puzzles here is not at all like the rating system on lichess or tempo. I honestly don't understand what determines a puzzle's rating here on com at all. The pass rate on some of the puzzles I have seen are way out of line. Say, for example, you have a puzzle with a 25% pass after 30000 plays (i just played one like this earlier this evening). Shouldn't the pass rate fall somewhere around 40%-60% if the puzzle is at a stable rating? Why doesn't the puzzle's rating move? It does on other sites. Why not here?

The rating system makes no sense? The rating system is bizarre?

Well, that's cos it's no rating such as the Elo rating that reflect the outcome of games played between rated players.

It's "puzzles", NOT chess games. Hence, it's maybe some sort of rating, but hey, it's not the Elo rating. I mean, not even anything close to it.

And frankly, these reward points chess.com grants to puzzle solvers, are purposedly looking like an Elo rating. It has to smell like chess. But it's only puzzles. Chess puzzles okay, but still puzzles.

Voilà.

It’s not a difficult puzzle when you play the moves by process of elimination. After white queens, white is threatening devastation with Rh8, so you know your move has to be a check. You only have two checks that don’t give away a rook, so it has to be a check on f4.

Between …Rgf4+ and …Ref4+, we can dismiss the second one as white’s King will run to the queenside.

so it’s not necessary to think ‘I’m looking for a mate in 3’ or whatever, we play the only move that makes any sense. On the second move, rinse and repeat (although move two is even easier as you’re reduced to ONE possible plausible move).

It’s a completely unremarkable puzzle and not necessary to ‘predict’ any of the opponents moves at all.

I'm saying it's weird because other sites do a puzzle calculation in much the same way elo and glicko work. Each puzzle's rating does move over time relative to the players who solve it based on their puzzle rating. Com is the oddball here.

They do change for a certain amount of time and then get locked to a level at a certain point. Staff will occasionally modify that based on things like solve rate if the rating seems to be off and it gets reported.

They use to change after every solve but changed to the above method at some point, as I understand it.

Is there any way to report a puzzle?

There used to be a ! button, but now it's gone. And even when it was there, there was a warning to only report puzzles that are wrong, not for trivial matters like a rating being 1000+ points off.