This would go down well on Reddit, where that style of puzzle is almost the only one appreciated. And it's a good way to quickly drill a few checkmate themes. But for chess improvement I would recommend a more general tactics trainer. IMO Chesstempo cracked the problem of automatically generating and rating interesting problems, but if you find the interface there too dated, lichess or chess.com have a reasonable second-best.
I had a similar experience to the person above - after a few puzzles where the first move is the same I wasn't interest in doing more. I also don't want a random puzzle - generally puzzle progression goes from easy to hard - will random just give me a really hard one?
If you are set on keeping the order of _1001 Ways to Checkmate_, maybe giving some context on which chapter the user is on would be nice (Queen Sac, Forks). Otherwise as a user, if I don't want to do some unknown amount of queen sac puzzles over and over then I'm just going to bounce.
How did you enter these positions into the app? Manually?
One idea you can try to avoid manual work:
First, download a game database, like for example a Lichess database,[1] or a database of Masters games. After parsing the PGN, check if the game ended in checkmate. If so, then at the very least you have a Mate in 1 puzzle. Using your engine, you can check the last couple positions of the game to see if there was a longer forced mate. If the game didn't end in a checkmate, it's still possible the resignation was due to forced mate, so you can do a similar process to check the final position.
By removing manual entry, you can add a lot more metadata that the user can search for to customize their training. For example, you could add the ability to train specifically on rook checkmates or knight checkmates, or queen sacrifices, or checkmates that occur in the back rank, or endgame checkmates (positions where there are few major pieces on the board).
https://database.lichess.org/#puzzles
> Generating these chess puzzles took more than 50 years of CPU time.
> To determine the rating, each attempt to solve is considered as a Glicko2 rated game
So you can filter problems for queen sacrifice or mate in 3 but not the combination of the two.
The second quote means that the puzzle rating is dynamic and changes every time someone tries to solve it (the puzzle gets points for losing / winning the same way the player wins/ loses rating for solving the puzzle. )
Were you doing web dev way back in the day or is this based on a template you modified? It reminds me a lot of a tool I once made. It was before I was programming professionally but I had a blog and was figuring out bit by bit how to customize things and animate them. Fun times!
The main thing I'd recommend is having illegal moves do nothing, and make drag and drop more clear. The "floating ghost piece" is actually not the cursor, and that's kinda not very nice UX. You might try a "click to place" technique instead of drag & drop.
Is this book out of copyright now?
> Is this book out of copyright now?
I think so. The copyright term back then was 28 years, and AFAIK the copyright wasn't renewed.
Really large set of curated problems, which can be divided into particular problem types (e.g. attaction or desperado tactics). Further, after you do the problem, community discussion forum of that problem becomes available, where they discuss alternate choices and their problems etc. It really is high quality.
There is another mode (mixed) which includes problems that are about defending.
