lichess.org
Donate

I figured out how to beat bots.

Make safe, positional moves. Do not try to complicate the position to make them blunder. They don't get confused by complexity. They blunder when a move counter tells them it is time to blunder. So play good, simple moves and watch for the blunder. Once you win a piece, watch out for their sneakiness.

I almost beat Antonio. I had to think about each move at least 30 seconds, some a minute or so. Normally I play like it is a 10 minute game. This time I made fewer mistakes and saw the dropped piece. But Antonio got me with a sneaky, non forced back rank checkmate.

That bot is not 1500. But I don't mind. Maybe as strong as a 1500 at classical time controls. I can beat the 1600 bot most of the time with far less effort. It likes going to endgames, and I'm pretty good at endgames.
Yeah, bots don't like positional games that much. They also are not a fan of the Hippo, a very slow and positional Opening. Maybe that is the reason why we "understand" open games so fast and have struggles with closed positions? I mean, if a database-computer struggles, we should struggle even more, right?
I just finished looking at the score sheet against Antonio. The piece I won was by removing a defender. It did not do a simple hang. Its other blunders were failing to execute 5 move combinations that would have won an exchange or a pawn. Most of my mistakes were failing to kick a piece off a good square, or allowing a knight to trade itself for my bishop. The ! moves were just forced recaptures, since failing to take back would be several points lower.

As for the mating threat, had I seen it, my response would have been a positional blundering, requiring me to retreat a doubled 7th rank rook to save my king. The engine said the best save was to give up a lawn in front of my king so no mate is possible.

I also noticed that often the best move alternative to a mistake I made was one I considered extensively but dismissed because I'd lose a pawn. But that pawn was going to be lost regardless.
@Chesserroo2 said in #1:
> Make safe, positional moves. Do not try to complicate the position to make them blunder. They don't get confused by complexity. They blunder when a move counter tells them it is time to blunder. So play good, simple moves and watch for the blunder. Once you win a piece, watch out for their sneakiness.
>
> I almost beat Antonio. I had to think about each move at least 30 seconds, some a minute or so. Normally I play like it is a 10 minute game. This time I made fewer mistakes and saw the dropped piece. But Antonio got me with a sneaky, non forced back rank checkmate.
>
> That bot is not 1500. But I don't mind. Maybe as strong as a 1500 at classical time controls. I can beat the 1600 bot most of the time with far less effort. It likes going to endgames, and I'm pretty good at endgames.
Get REady to beat stockfish.
@MasterOvFan said in #4:
> Get REady to beat stockfish.

Stockfish is not programmed to make bad moves periodically. To beat stockfish, adding complexity might not hurt.
@Chesserroo2 said in #1:
[...]
> That bot is not 1500. But I don't mind. Maybe as strong as a 1500 at classical time controls. I can beat the 1600 bot most of the time with far less effort. It likes going to endgames, and I'm pretty good at endgames.

Not everybody knows who "Antonio" is - a little bit more context would be nice...

So... For all those, who don't know which bots we are takling about => It's the bots on chess.com (play a game against the computer).

Imo the bot rating numbers are totally nonsense and can't really be compared to "real life" (the same is true for lichess, too btw.).
What i can say from my own experience is that the bot Li (showing a rating of 2000) must be definitely much less than 1500 ELO FIDE, otherwise i wouldn't win 8 out of 10 matches against it.

Have fun!
Computers generally do computer things really well and humans do human things really well
To win a bot you must buy a bot, get it?

So winning a bot (means winning something) you can taks part of the lottery.

Logic.

This topic has been archived and can no longer be replied to.