Training Problems

@Logando2805 good puzzle, I just looked for forcing moves and it worked. 10% pass rate is a bit low though.
@Valesias120 idk, is it Rxf7 and Bg2 ?

I think the puzzle is technically impossible because you don't know if Black can still castle or not. The move it wants is 1. Rd1, which I considered, but after 1... O-O Black is just better. So, I played Rxf3. When I analyzed the puzzle with the engine it said that Black was unable to castle. If Black can castle Rxf3 is the best move.

aah I see what you mean. Honestly, I didn't even consider castling for black 😅

It is indeed! The quiet backwards bishop move sets up an unavoidable loss of material.

I can't find the solution. What do you think guys ?

Nd4+ if Kb2 Qd2+ Ka1 Qxc3+ Rb2 Bxa3

if Ka4 Qxc3 and white loses material

@DividingCart27 beat me to it! Nd4+ looks like the best move to me. Na5+ was my other big consideration, but after Ka4 I think White starts to escape the mating net.

I find it hard to visualise after Nd4+ Ka4 Qxc3 why it's good for black. Don't forget that white is a rook up. maybe I am tired for now

I looked at Na5+ first but quickly realised it didn't work after Ka4 Bxa3 Kxa5 and there is nothing. When I saw Nd4+ I just knew it has to be the move knowing it's a puzzle

You are threatening to take his knight with your bishop leaving his bishop defenseless

The important thing to realize before you even start calculating these puzzles (IMO) is to see if your opponent has an immediate threat. Since White doesn't have any good checks, Black has time to play a quiet move. Qxc3 is "easy" to find because it is a forcing move, threatening Qxa3#. Once you see the threat of Qxa3# you have to see that none of the moves with the Knight are good.

It's a bit of a brute force calculation, but that was my thought process. That being said, there is a lot of variance with puzzles. I've definitely missed way easier puzzles than the one you posted.

I'm answering late, but I just remembered the rule about puzzles: If the opponents' King and Rook are in their starting position and it hasn't been explicitly said that castling isn't possible, then it is possible. But in this case it wasn't possible, yet this is not stated, so it is a mistake on their behalf.

I mean yeah ofc, Qxc3 is game over for white. My mind was off yesterday, I couldn't see after Qxc3 Rc3

Here is a great calculation exercise from today's round of the FIDE Grand Swiss 2025. The game is Abhimanyu Mishra v Gukesh Dommaraju. White has only one winning move. Enjoy!

Find the best move for black (advanced/master). Position from https://www.chess.com/analysis/game/live/143025128348/analysis?move=46

No way a human can spot this without spending a ton of time. Is the idea that White is in a Zugzwang after h5?

It just a better endgame I think. All the lines are in the exercise so you can compare the endgames. The bishop looks badly placed on d3.

Find the best move for black (advanced/master). Interesting position from one of the top board of the FIDE Grand Swiss move 43... (https://www.chess.com/events/2025-fide-grand-swiss-open/08/Keymer_Vincent-Vidit_Santosh_Gujrathi)

Yeah, I watched it live. When Vidit played c6, Vincent couldn’t believe it, he instantly recognized it as a blunder. Check his reaction : 
https://youtu.be/ZM9FfeK0-P4?t=17110

Find the best move for white and try for eight minutes before reading more (intermediate/advanced). After thinking for eight minutes I had to look on the hints that was Mating net and Zwischenzug. I was also reading the instruction that helped me find the solution. "Black has a weakened king. White's queen and bishops are looking in his direction. But the piece on d6 is under attack. Retreats should be evaluated last. Look for an active game plan first!" After thinking for maybe twelve more minutes I found the solution that was worth all the effort. Problem from Day 4 in the 50-day tactics challenge Part 2 course (https://www.chessable.com/50-day-tactics-challenge-part-2/course/262474/). Full explanations in the course. (position from Levon Babujian – Jakov Geller (2008))