I've noticed all of the bots have certain opening lines prepared and they are most certainly integrated and the bot isn't "finding them" in game so to say. Play the same moves against the bots and they will always repeat the same moves.
Isabel-bot (1600 rating) playing a GM game from 1995
Sort:
Mar 8, 2021
0
#3
I think most of the bots are adaptive: if you play very well for a set of moves, they tend to do the same. I'm sure Nelson do this, I noticed analyzing the match: everytime i made a mistake, he made the same kind of mistake 1 or 2 moves later. And usually I win against bots when i start to play almost random and then suddenly begin to do my best.
Came here to say this. No way Isabel is 1600. I trounce the 1700 engines regularly and rarely beat this one [Isabel].
Jul 10, 2022
0
#5
What I have noticed is at times the bots seem to be much lower skilled than their ratings would suggest and at other times, they seem much higher. I think this is deliberate to give us the sense that we are playing real games against people instead of just against the sgm engines that are toning down their game (which is I think what is happening.)
It is frustrating to see the bot go from very bad to very good and not understand why. I don't think it is a fun experience. However, it might be beneficial as a training tool.
Also been trounced by just this bot at that level.
Mar 3, 2023
0
#7
All bots are modeled to copy moves of games for the first little while. I play bots with weird openings. That makes them move on calculation instead of lines
Jan 7, 2024
0
#8
This bot is without a doubt an outlier. I set a chess engine to 1600 and beat it eventually with relative ease, but Isabel was a nightmare. The only thing that worked out for me was relentlessly checking the king and then she seemed to drop the whole 2000+ rating gameplay. Seen several accounts of other people having the same experience.
Then I moved on to the next bot at 1800, and his moves were way less frustrating and more enjoyable to play against, because it actually felt like I was playing the bot at the described level. I could feel that all my moves didn't have to be millimeters away from perfect without getting beat down entirely, which is usually the case with Isabel. Something is seriously wrong with that bot and they should do something about it.. Intermediate players going over to their first set of advanced plays are inevitably getting so discouraged and downtrodden by suddenly facing this monster..
Hi everyone.
Quick question here. I was practicing my chess against Isabel-bot and while using the 365chess database on the side to analyse the best moves, I can conclude without any doubt that the bot uses the best moves possible for black for the first 15 first moves following the exact same moves from a game between GM Karpov (2780) and Anand (2715) in 1995.
How can she be 1600 rating? She's playing like a GM...
She's suppose to be good in end games, but she knows by heart the best move possible for black for 15 moves, while completely losing the late game instead of getting a draw as it happened in real life?
Make no sense to me at all.
It only happens to me when she uses the Caro-Kahn defence though.