Though this is a very neatly made and powerful script, I want to give some helpful pros & cons. - Your UI layout and styling is 10/10: It's hard to put into words how impressive the styling is for the limited amount of options you have to work with, good job. - Your "humanization" technique is very strong: The only issue, however, is that Chess.com's cheat detection is significantly more advanced than what your script sums it up to be. Chess.com goes down to the level of genuine human moves, not just moves that aren't the top engine picks. This includes things such as playing move SEQUENCES faster, playing a series of moves that are indicative that you don't know what to do, or moves you find that aren't necessarily good moves, but you still play rational follow-up moves that lead behind a deeper idea other than just the move being bad against a perfectly playing engine that wouldn't let you do anything. - The board play is pretty inadequate: The stockfish.js file isn't the best option you have for an engine (Stockfish 10.0.2 is the best local model, but it's pretty much garbage; even 2000-rated players can beat it in rapid games). Stockfish 10.0.2 is good for being a backup local model, but it's only viable in bullet games because you either have to wait .1 seconds for a pretty good move, or a full minute for a good move (the model handles depth increases horribly). I would use an API-based model, such as the one at https://chess-api.com, which is the one I use for Stockfish 18.1 with pretty good response times (although it is used so much that it's on cooldown some of the time). Also, from just a script perspective, I don't think the auto-play works at all. I tested it numerous times, and it never ended up actually moving a piece on its own. Not only this, but the engine is set, by default and unchangeable, to a depth that is too high for the Stockfish 10.0.2 model to produce a move within a short period of time (should be less than a second at maximum). There are also some minor layout kinks with the move display and the other visuals, but I'll leave that up to you to test and debug.
Pro tip: Have all the settings save locally through the userscript so that the user doesn't lose any of their settings when they reload or come back to it later.
Hi rexxxC#,
Though this is a very neatly made and powerful script, I want to give some helpful pros & cons.
- Your UI layout and styling is 10/10: It's hard to put into words how impressive the styling is for the limited amount of options you have to work with, good job.
- Your "humanization" technique is very strong: The only issue, however, is that Chess.com's cheat detection is significantly more advanced than what your script sums it up to be. Chess.com goes down to the level of genuine human moves, not just moves that aren't the top engine picks. This includes things such as playing move SEQUENCES faster, playing a series of moves that are indicative that you don't know what to do, or moves you find that aren't necessarily good moves, but you still play rational follow-up moves that lead behind a deeper idea other than just the move being bad against a perfectly playing engine that wouldn't let you do anything.
- The board play is pretty inadequate: The stockfish.js file isn't the best option you have for an engine (Stockfish 10.0.2 is the best local model, but it's pretty much garbage; even 2000-rated players can beat it in rapid games). Stockfish 10.0.2 is good for being a backup local model, but it's only viable in bullet games because you either have to wait .1 seconds for a pretty good move, or a full minute for a good move (the model handles depth increases horribly). I would use an API-based model, such as the one at https://chess-api.com, which is the one I use for Stockfish 18.1 with pretty good response times (although it is used so much that it's on cooldown some of the time). Also, from just a script perspective, I don't think the auto-play works at all. I tested it numerous times, and it never ended up actually moving a piece on its own. Not only this, but the engine is set, by default and unchangeable, to a depth that is too high for the Stockfish 10.0.2 model to produce a move within a short period of time (should be less than a second at maximum). There are also some minor layout kinks with the move display and the other visuals, but I'll leave that up to you to test and debug.
Pro tip: Have all the settings save locally through the userscript so that the user doesn't lose any of their settings when they reload or come back to it later.
Good luck,
Ech0