A More Realistic Type of Puzzle

Over the years I must have done thousands of chess puzzles, both easy and hard. However, one thing that has always bothered me about these puzles was that you knew there was a solution. After a while I would just start to look at ways to sacrifice my queen, and if that didn't work I would sacrifice the next strongest available piece. Usually this technique was enough to checkmate or win material and I would quickly move on to the next puzzle.

That's why I enjoy the following position, taken from a  game of mine from 2012. I had blown a winning position on the black side of Evans Gambit, and my opponent, with just a couple of minutes on his clock had a choice of appealing ways to win material. However, only one of them works. Can you find it?

Now for what actuall happened...

Did you find the win, or settle for the first move that won material?

This is why I emphasize to my students to always look for the forcing moves for both sides. If there's still action on the board, it's too early to end your calculation.