I need help understanding this 2600 rating puzzle

I got this puzzle while attempting Puzzle Rush Survival today. I correctly calculated the solution, but didn't see enough of an advantage with it and so I played my line (incorrect) because I believed it was winning (it was winning, but not as good as the puzzle solution?)

My annotation comments explain my thought-process. Any insight on why the final position (Rxc1) in this puzzle is over +2.00 evaluation, I'd be interested in. 

p.s. for reference, this is puzzle: https://www.chess.com/puzzles/problem/2084238 

This puzzle is way above my level, but it is a good question.

Hello Keseto. I think the final evaluation is very much in white's favor (after 21. c5) because black will not be able to do much to prevent the knights coming to c3 and d4, followed by a knight to b5 and coming in the d6 hole. a6 at any point to prevent this would be bad because it creates a hook for a pawn storm on the black king. Now black as nothing left to do than suffer with white knights placed at d6 and f6 or g5. I think to reason why your other line, wich looks very promising too, isn't as highly regarded by the engine is because black is keeping some minors pieces to trade off white knights once they land on their natural outpost. Probably that the engine will refute everything i just said but from a quick glance at your position i think it makes some sense

Thank you @exterminator444 I figured it would be a positional advantage like this where black would be just helpless to a gradual downfall, but I still am not sure why the puzzle solution is "best" and only about half a pawn better than my line.

I was going to report this puzzle if some decisive conclusion doesn't come out of this forum thread. If one line was winning by half a pawn and the other was equal, then I maybe could justify the puzzle solution, but here both lines appear winning (with little difference) and I still feel confident with my line being +1.50 and being able to convert that game into a win with white most of the time.

I toyed a little with the engine. It seems that in the engine's line the following Nfd4 is a monster here

Ooh very nice @introuble2

I didn't mention it, but I also figured that Nfd4 would be a good move because I was thinking of Jeremy Silman's Imbalances and also it is nice to blockade the d5 pawn. I didn't see the move c7+! though with the incredible geometry which makes this a killer move though

The knights are simply too strong here, I agree with the others. If black plays a6 then the b6 square is very weak as well.

yeah i think c5 leads to a much bigger advantage, but it is much more of a positional idea than a tactical one. overall i just think the puzzle is low quality

I personally have a love hate relationship with this puzzle. I love it because it's more positional and can be in conjunction with Jeremy Silman's imbalances, which is my favorite book and already mentioned by you. However, I hate puzzles where the top 3 solutions are all super similar, as in this one, and you get punished for not playing the absolute one. Alas, chess.com's puzzles are pretty good, so I'll accept the occasional annoyance for the overall general quality.

For one, both positions are likely winners for white - and I agree with that. A tactical player would prefer the position with the queens on the board and launch a queen side attack. But that is not the relevant question. The problem is that the evaluation scores are so close (about .8 points apart) that it is unreasonable to expect a human player to be able to choose between them in a puzzle. Commonly the point gap between the best and the second best move is like 2 points or more but this one is simply unreasonably small. So I think chess.com has some explaining to do.

Possibly the evaluation scores change suddenly with another 10 minutes of computation time but the puzzle fitness remains debatable when separating lines takes such a long analysis. Perhaps the engine has a change of heart after finding a significant improvement on depth 99. But how can you expect a human solver to detect the same discontinuity on such an extreme depth? It stands to reason to expect that the puzzles are not engine-analyzed for longer than a minute or so!

Yeah, that's not the best puzzle to be honest; when there are two lines that are both giving an advantage, the better of the two good lines should be overwhelming or at least very easily demonstrable as to why it should be preferred. Obviously in the solution line white is a lot better but it's not 'resolved' in the way you expect puzzles to be.

On the note of c5 you want the position closed because now that they used up their attackers black dosnt have the assets to maintain their position while white does.

says something not related to chessw ithout even reading the post epicly

Getting into puzzles of this level usually kicks my tail - especially if anything positional over material. Somehow on this (following your link before reading further) one though, I was debating between what turned out to be your solution and the "correct" one. Almost went your direction, but settled on the latter, and accidentally got it "right.". Thanks for sharing an interesting puzzle, even -or especially- with the conundrum of a nearly equal alternative solution.

bummer right

Back when I was 1800-1900 range (about 2 years ago), I used to suffer from this flaw a lot; I would calculate really hard on the line I want to work and then when it isn't satisfactory, I would just play the "other move" with little thought. Sometimes it worked, but a lot of the time it would be a silly blunder and especially in real games where there is no guaranteed "solution." I'd like to think I've moved past that faulty mindset since then.

As for this puzzle, I calculated both (as my annotation comments describe), but your insight is valued llama and you are absolutely correct about me misevaluating the knights as that much better than the h8 rook, but once pointed out, now it makes perfect sense!

No worries; I sometimes miss things too and especially when tired.