- Blind mode tutorial
lichess.org
Donate

GM Duda smarter than a computer analysis

#1

In a recent game against Anish Girl, Jan-Krzysztof Duda made the move 28. Rg7+. The full-game computer analysis on Lichess evaluates it as a blunder. However, a deeper cloud analysis on Lichess (conducted due to the tournament's prominence) reveals it to be the only move for a mate-in-11 -- quite the opposite of a blunder. Duda found the mating sequence and several moves later Girl resigned. The difference in evaluation isn't due to any bugs in the code but simply the different depths to which the position was analyzed.

I found this striking discrepancy to be another useful reminder that while the computer analysis wonderfully offered for free (with some limits) on Lichess and its results can be quite useful, it is just a single tool and should not be seen as absolute truth. So take care in using it. As far as I know, Lichess has no plans to release a "Duda, analyze my game" feature .

https://lichess.org/broadcast/meltwater-champions-chess-tour-finals-2022/round-3/CTOuD9xq/ZmaTKyA3#54

In a recent game against Anish Girl, Jan-Krzysztof Duda made the move 28. Rg7+. The full-game computer analysis on Lichess evaluates it as a blunder. However, a deeper cloud analysis on Lichess (conducted due to the tournament's prominence) reveals it to be the only move for a mate-in-11 -- quite the opposite of a blunder. Duda found the mating sequence and several moves later Girl resigned. The difference in evaluation isn't due to any bugs in the code but simply the different depths to which the position was analyzed. I found this striking discrepancy to be another useful reminder that while the computer analysis wonderfully offered for free (with some limits) on Lichess and its results can be quite useful, it is just a single tool and should not be seen as absolute truth. So take care in using it. As far as I know, Lichess has no plans to release a "Duda, analyze my game" feature . https://lichess.org/broadcast/meltwater-champions-chess-tour-finals-2022/round-3/CTOuD9xq/ZmaTKyA3#54
#2

Just to reinforce the point, I gave the position to a Stockfish 15 mobile app. It spent merely a few seconds on Rh7 before finding the forced mate.

Variations in resources, speed, allotted time, & hardware all contribute to the accuracy of analysis.

Just to reinforce the point, I gave the position to a Stockfish 15 mobile app. It spent merely a few seconds on Rh7 before finding the forced mate. Variations in resources, speed, allotted time, & hardware all contribute to the accuracy of analysis.
#3

@Davelz said in #2:

Just to reinforce the point, I gave the position to a Stockfish 15 mobile app. It spent merely a few seconds on Rh7 before finding the forced mate.

Variations in resources, speed, allotted time, & hardware all contribute to the accuracy of analysis.

Nice. I suspected somebody would find something like this. Remarkable nonetheless.

@Davelz said in #2: > Just to reinforce the point, I gave the position to a Stockfish 15 mobile app. It spent merely a few seconds on Rh7 before finding the forced mate. > > Variations in resources, speed, allotted time, & hardware all contribute to the accuracy of analysis. Nice. I suspected somebody would find something like this. Remarkable nonetheless.
#4

I guess in some form or another humans still are able to keep up.

I guess in some form or another humans still are able to keep up.
#5

It all depends on the point of view. If a human grandmaster analyses a position, he will have a different opinion than an engine to grandmaster level.

I'll just keep playing my dearly aggressive openings.

It all depends on the point of view. If a human grandmaster analyses a position, he will have a different opinion than an engine to grandmaster level. I'll just keep playing my dearly aggressive openings.
#6

I can't fathom the analysis skill required for Duda to do this. Quite incredible.

It reminds me of a story about Kasparov at his peak, when early in a game he took forty minutes over a move. Many moves later, the game was over after a forced sequence of a ridiculous amount of moves leading to checkmate. When he was interviewed afterwards he stated that he had calculated not only that line through to the end, but also - and you'll forgive me here for not recalling the exact amount - another eight lines or so. Unfortunately I can't recall the game, which goes to show why he is Garry Kasparov and I'm not.

I can't fathom the analysis skill required for Duda to do this. Quite incredible. It reminds me of a story about Kasparov at his peak, when early in a game he took forty minutes over a move. Many moves later, the game was over after a forced sequence of a ridiculous amount of moves leading to checkmate. When he was interviewed afterwards he stated that he had calculated not only that line through to the end, but also - and you'll forgive me here for not recalling the exact amount - another eight lines or so. Unfortunately I can't recall the game, which goes to show why he is Garry Kasparov and I'm not.
#7

Le Quang Liem is proving himself to be smarter than a human who is "alleged" to be smarter than a computer.

Le Quang Liem is proving himself to be smarter than a human who is "alleged" to be smarter than a computer.
#8

@george_mcgeorge said in #4:

I guess in some form or another humans still are able to keep up.

As most gloriously shown in ZugAddict winning a special Lichess trophy for defeating Stockfish.

@george_mcgeorge said in #4: > I guess in some form or another humans still are able to keep up. As most gloriously shown in ZugAddict winning a special Lichess trophy for defeating Stockfish.

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