A mini-revolt against computers in chess

Mikhalchishin is not an advocate of too much computer use. ‘Engines like Rybka, although very strong, can be also very dangerous, because after an hour of a computer analysis the player is completely under the Rybka’s guidance and can’t invent anything, just follow the machine. They can analyse some position, but it is very difficult to get a valuation of a position with Rybka – there is always something unclear, you never know what the real variation is. Rybka takes a lot of mental energy. Computer analysis switches off the brain. I enjoy seeing how the brain works, not computers.’

There is more here, for instance:

…he feels that an interesting trend is taking place in the chess world presently: a new generation of players, that he calls ‘post-Carlsen generation’, is coming up; young players who are not so much dependent on computers and are more practical, ‘hand players’. Carlsen may even become a world champion, but at this moment, a new generation is growing and training. ‘Richárd is one of them; then there is Nyzhnyk, a very interesting player from Ukraine, Berbatov, a very talented young player from Bulgaria. But the leader of this generation I would say is Wesley So. He is extremely talented and has produced some very interesting games, like his wins against Ivanchuk at the World Cup. These post-Carlsen players have a different style and attitudes. They are not obsessed with the opening theory, like their older predecessors. They are looking for much more practical play and are very aggressive. They are not necessarily a computer generation, as Carlsen’s generation was. Computers came with their powerful programs and chess players wanted to try them. But I feel this trend is finishing now.’

I wouldn't put too much stock in this as a practical development (Carlsen's the guy who's #1), but it's an interesting point about the roots of creativity and independent thought.


This sounds to me like wishful thinking. The sad truth is that as the gap between computers and humans widens, human chess competitions will probably lose appeal increasingly, possibly to the point where they'll seem as pointless as, say, arithmetic competitions. Thus, it's not surprising that many chess enthusiasts dream how nice it would be if computers would just go away.

Another mortal threat to high level competitive chess is that with increasingly powerful and miniature computers and communications devices, I don't see how it will be possible to prevent cheating in the foreseeable future.

That's why Kirk can beat Spock in chess.

One problem is that the annotation/analysis features in computer chess suck. The publically known state of the art (i.e. not considering Chessbase's proprietary engines and the like; they do have an 'annotation' feature but for all we know it's pretty much a gimmick) is limited to outputing a principal variation (i.e. the suggested move and the resulting minimax play) and a numerical evaluation of the final position.

You can do better by looking at iteratively deepened analyses (this might tell you whether the move is a routine or surprising one), considering alternate moves at various points in the variation (which clarifies whether these moves are forced and provides alternate paths) and perhaps looking at branching factors and tree depths (is the position a tactical or strategic one?). But some of this can be computationally expensive, and the result is still very middling compared to human analysis. There's clearly a vast scope for research here.

Comments for this post are closed