Zero rated puzzles or zero rated raters?

I often come across zero rated puzzles which I think are quite good but are simply not understood. For instance:

lichess.org/training/69149

In this puzzle, black seems to have a choice between saving his knight but letting white's bishop escape, or taking the bishop and giving up the knight (after which each side can capture the guilty pawn). If that were all that was going on, the puzzle moves would be incomprehensible. Moreover, if that were all that was going on, it might be preferable to take the bishop, although that would hardly merit a puzzle. However clearly there is more going on. Black also has a mate threat with the queen in the corner that can be defended against by the white knight (after Qa1+). However, by saving black's knight black is also putting the knight in a position to move to c5 and help the queen mate on b2. There is no way white can save his bishop and defend against the mate threat, but it takes some time to see that. Don't tell me - you zero raters - that you saw all that and still thought the puzzle sucked. You simply didn't understand why the computer moves white's king and gives up the bishop. The computer "sees" the mate threat that the raters do not. Admittedly one may stumble on the solution without having fully seen the implications of saving the knight. That may detract slightly from the value of the puzzle, but only slightly. I think many of the zero rated puzzles are like that; the raters don't really get it.

that's why after you get it wrong/right, they conveniently load the engine so you can actually see why the move you "expected" was so trash. It is a bit annoying, but that's life in automated systems.

Actually i solved this a little ago (now at 69530), but i rarely vote on any puzzle.

I liked it:)

One has to take the rating of the puzzle with a grain of salt. Perhaps people should be more modest before rating a puzzle. If the puzzle didn't make sense to you, or if it seemed incredibly simple, or the engine move seemed dumb, consider that maybe you didn't get something. You can discuss it here before immediately taking out your frustration on the puzzle rating.

You don't have to see all this to get it right. The concept of this puzzle is simple.

If you take the bishop, white can take the knight with tempo and save a2. If you move the knight the bishop and the pawn a2 is hanging. So you left white with two problems.

That's true, and no doubt why people stumble on the right answer without seeing the mate threat. However, as I said in the original post, after fxg4 exd7, black would take the pawn on d7, eliminating an annoying advanced central pawn. Meanwhile white saves a2 only by advancing it to a3 where it may soon be targeted by black's bishop (after a3, Ng3 opening the diagonal). You're right that that is nevertheless much better (or less bad) for white, but especially because Qxa2 is a potential mate threat when coupled with the knight. The fact that white has two problems is inherent in the puzzle to begin with, with black to move first. The engine recognises the gravity of the a2 threat when coupled with black's N to the point where it prefers saving it at the cost of the bishop. People who don't see that may make the right move without realising that it wins the bishop.

Sorry, I should have said that white can also move the king to b1, after which black will have a harder time mounting a serious threat with the bishop.

"One has to take the rating of the puzzle with a grain of salt" he said as he wrote an essay complaining about puzzle ratings.

"One has to take the rating of the puzzle with a grain of salt" he said as he wrote an essay complaining about puzzle ratings.

Where's the contradiction? The "essay", actually a short paragraph, is explaining why the puzzle ratings are unreliable and therefore should be taken with a grain of salt. What problem do you have with that? Some people have written whole books arguing that the bible is an unreliable guide to morality or history. Do you think that's contradictory too?

Here's another zero rated puzzle which I think is fine:

lichess.org/training/69203

Again, probably the low rating derives from the ease of stumbling on the right moves without having seen all the steps in advance; however it takes thought to figure out that one either has a mate threat or wins a piece.