To get ready for my first tournament OTB, I thought of playing some rapid training matches. Under USCF rules for playing a match the difference between the two players cannot be over 400 points. A rapid match would have as longest time control: G25 d3 (game in 25 minutes with delay 3 seconds) anything above will become dual rated, for example a G30 D5 would be rapid and long time control for USCF.
Rapid games are really good to gauge the understanding of a new opening repertoire, or see the parts where one should study more. With that in mind I tried to contact some of the chess “friends” who could be interested, and many are only interested in long time control games, which for me at the moment are not worth it, because I would need to find players higher rated than me, and it’s not so easy, or have that kind of time and I don’t. Some were likely too afraid, and some have the wrong mindset: they believe that rapid games will not improve their skills and chess understanding.
Online unfortunately cheating is more than pandemic, chess-com closes about 35000 accounts a month for cheating, and those are a small part of the top of the iceberg, which gets caught.
I decided to try to play against one of the many chess-com bots.
I remember many years ago, there was a software called Chessmaster which had different personalities, and they were quite good. I thought 10 years later, they should have improved, but it wasn’t the case, and this match will prove it.
I chose from the advanced section
and I choose Lorenzo, partly because he’s Italian, he’s an 1800, so it should give me some troubles, but maybe not too much, and definitely spot my weaknesses. For USCF quick rating I’m around 1700, so to choose a player 100 points above me, should be a good training.
I played 6 games, 3 with White and 3 with Black, and the result was a stunning 6-0!!
Lorenzo didn’t play like an 1800, he was more of a 1000-1200, blundering pawns and pieces, and losing control of the game when one begins to attack sacking pieces.
I cannot understand how chess.com gave these bots a rating. I am also pretty sure that against a human 1800 I would lose at least 1-2 games and draw for sure 1 or 2.
But the real question is: why after so many year Chessmaster was released, nobody developed engines with human like play? As a friend of mine at master level told me, to play against an engine at full strength is depressing, because it will smash you within 25 moves, or gain an advantage which is enough to win the game.
The games are annotated lightly, but as you can see this bot blunders in a way who even a 1500 rated player would win.
There are also many other functions missing. I couldn’t get a time control like for real OTB games. The games were unrated. I played them using my Ipad pro, because I made the moves and played the game on a real chess board.
Finally here the games:
Final Thoughts: While analyzing the games I felt this match was useless. Generally when playing against humans I calculate many more lines. I remember them after the game and write them down in my annotations. Unfortunately these six games didn’t help me like six games played against another human would have.