Puzzles

@tiamat6192
for puzzle #8Cifm, after Bxf6 by black, Stockfish at depth = 23 says Qxf6 is best, giving it a -4.9 evaluation. The 2 next best moves (Qf4 and Qb4), are evaluated at -7.3 and -7.4 respectively. Now, the evaluation bar is ranked so that the higher the evaluation, the better the winning chances are for white. +0.3 would mean having a small advantage for white, -1.6 might mean being up a pawn for black, 0.00 means the position is equal, and so on. This means that if you are playing black, you would like the evaluation to be as low as possible, and white would like to have the evaluation as high as possible. So, if you were playing white, since -4.9 is greater than -7.3 (both numbers are negative, but one is still bigger than the other), you would want to choose Qxf6.
As for your question about attacking, if white moves the queen to g3 or b4 or f4, looking ahead on the analysis shows that the attack is based around targeting the pawns on b2 and c2, like a pawn storm on the king with the c6, b7, and a7 pawns.

Puzzles are usually never wrong, so don't assume every puzzle you do has 10 different solutions, just try to find the best move and solve it.

@thibault
"@adwait_garg "there might be many ways to solve a particular puzzle"
Absolutely not. Show us an example of that."

I'm afraid you're the one who is wrong this time thibault ;)

This puzzle for example:
lichess.org/training/i6hpL
On move 47 in this puzzle only bxc3 is accepted, but both kxc3 and b3 are winning if you check with the tablebase:
lichess.org/analysis/8/8/8/P2p4/1pk1p3/2P1K3/8/8_b_-_-_0_47

Now that I might have gotten your attention, I'm curious: Is it too hard to code or too much computer power to check all (endgame) puzzles with the tablebase to avoid these things?
I recently got a puzzle (now removed) which could have ended out in a KQ vs KR endgame which obviously is a win, but the puzzle didn't approve of that solution. I think KQ vs KR is a common issue in the puzzles on lichess :(

this thread is a shit show lol

Wow, 63 TDs for the OP!

I think that the puzzle on #43 is the answer of "Puzzles never reject a move leading to a winning position":

But stockfish (+13, NNUE, Depth 26/26) changes the advantage that black has, when is played a different move of bxc3.

Before bxc3, Stockfish says that black has an advantage of -22.4 "pawns", after bxc3, Stockfish says that black has an advantage of -52.7 "pawns"

When is played b3, Stockfish says that black has an advantage of -9.5 "pawns". Nextly, after a6, Stockfish says that black has an advantage of -10.7 "pawns". Then, after b2, black has an advantage of -18 "pawns".

But, if is played Kxc3, Stockfish calculates that black has the advantage of -4.3 "pawns". a6 is the best (for white), so if played a6, black has the advantage of -3.6 "pawns".

My conclusion is that the puzzle has "a type of bug or glitch" because it doesn't accept the other 2 variants, but bxc3 is the best, b3 is good, but Kxc3 is the worst.

I think that the puzzle is glitched (because it doesn't accept the other 2 variants) and at the same time, that is fine (because bxc3 is the best, and it has logic).

Do I have errors in my conclusion? If so, could someone correct me?

(PD: I put the advantage that Stockfish says in "pawns" (with quotes) because that's how I learned what means the number that shows).

David.

Well, Lichess doesn't allows multiple solutions and that's why in a cooked (having multiple keys[first move of solution]) puzzle, even a correct move is stated as wrong.
I have seen such puzzles at Lichess but I can't share one as I haven't seen for few months now.

Wow, 70 dislikes for @tiamat6192 ! I've never seen this !