Blunderbot: the reverse chess bot

Welcome to my first chess.com blog post! In this one, I will discuss Blunderbot: a reverse chess bot.  As the name suggests Blunderbot isn't too good at playing chess, actually it plays the worst moves possible.

How it works

Blunderbot is based on a strong chess engine - in this case Stockfish. At each turn, Stockfish looks at all the possible moves and ranks them, and picks the worst one. If a mate is possible, it will ignore all other moves and plays the one that will give the shortest mate.

Games

Enough with the nonsense, this is probably why you're here for. For each of the games I will be using this configuration of Blunderbot:

• Based on Stockfish 15.1
• Time limit (per turn) - 10s

Blunderbot vs. Martin

Here Martin managed to mate Blunderbot in 28 moves. Blunderbot managed to make 2 blunders, 4 mistakes, and 11 inaccuracies. It also made 2 good moves and 8 excellent moves.

Hopefully, that made Martin a little better for himself.

Blunderbot vs. Blunderbot

Here I made some changes to the code that would make it even possible for the two bots to play against each other. The game took a lengthy 218 moves, ending in a stalemate.

The chess.com review is crazy:

Doing every move type in the game, totaly there are 20 inaccuracies, 41 mistakes, 90 blunders, and 81 missed wins. Anyone wondering about the brilliant move, it's 107. a7.

Blunderbot vs. Maximum

Here Blunderbot lasted a respectable 13 moves, especially knowing what it's designed for.

Conclusion

I hope you enjoyed a first look into Blundebot. You can find the source code here. Should Blunderbot play against someone else? Can I improve Blunderbot more? If you want to suggest something, you can write a comment. If you liked this post, sharing it will help me a lot. Thanks for reading!