When I was a child I occasionally played and enjoyed chess. Of course a few kids took the game more seriously for various reasons and so when I played these kids they would win and crow extensively over their victories. This was kind of annoying. No one likes losing, but grace in victory makes the loss easier to bear.
Even as a kid I just didn’t understand the value of winning a game where there was no real world gain for a victory. Winning a game because you had followed someone else’s instructions (from a book or personal instruction) seemed a bit hollow. What I did understand however was that having someone practically go into paroxysms of ecstasy because they had beaten me with the assistance of some cheap prepared opening trap was frustrating. I therefore resolved to learn more about the game and stop these displays.
It is the nature of chess that if you play enough you tend to get better. By the time I reached 16 I was top board in the school chess team and had won a few minor regional competitions – all of it sort of by accident.
For various reasons this was my chess playing zenith. Other commitments meant that I did not play with regularity although I did make a small effort to address this by purchasing what was then a top of the line chess computer. Over the next few years even this fell into disuse due to a working week that was in excess of seventy hours. Basically when I came home, intellectual exercise was near the bottom of my list of things to do.
Well, all this is over 20 years ago. My memories of chess are positive and so I recently decided on impulse to buy a chess engine from Amazon while browsing. Oh my! How chess engines have advanced! If I played against the engine on an unrestricted setting it would slowly crush me over and over again. Playing against it was such a shock I decided to go on the internet and research them.
Chess players in general are rated using the Elo system and it turned out that the engine I was trying to defeat had a rating (unofficial) of over 2900. To put this in perspective, the current world champion is Viswanathan Anand and he is rated around the 2800 region (giving a precise rating is pointless as it changes frequently). Most top Chess Grand Masters (GM’s) have ratings in the 2800 range and certainly over 2600.
This monstrous playing strength of course explained my difficulties in defeating my new found chess companion. To make matters worse a short search on Google showed me that not only could I have downloaded a free and legal engine that was a stronger player than the one I had purchased (so much for impulse shopping!) but that these newer engines were playing in the region of 3200 Elo. This Elo rating is truly amazing.
To help a none-chess player understand how impressive these ratings are:
A player facing an opponent a mere 50 Elo stronger might only be expected to win 43% of games.
A player facing an opponent 200 Elo stronger can expect to win a mere 25% of games
And finally,
A player facing an opponent 400+ Elo stronger can only expect to win roughly 8% of games played.
Remember a gap of 400+ Elo is a decent rough estimate of the current gap between the World Chess Champion (human) and the top chess engines.
So what is the point of all this? Well, for a guy interested in the future implications of computer development and the human role in this process the vein is a rich one, especially in the area of near future SF writing.
For instance: Can you imagine competing for a job in a cerebral environment when the best person in the world in that particular field can be out thought by a computer better than one in ten times?
Now it is true that there are those that will argue that chess is an extremely limited arena and is therefore easy for a programmer to deal with rules and variables etc. and thereby use the computers undisputed ability to number crunch and so extract a victory. “Real life is different” so these folk say.
This argument can be basically summed up by stating that the world is massively more complex and the tasks required to live in it more diverse and so computer intelligence is no threat to humans except in limited areas of competition where the environment is simple and lends itself to reduction to a simple heuristic. Further, that if computers do become better than humans in these simplistic areas then this can only be a good thing as it will relieve us of the boredom and drudgery of having to complete these tasks ourselves.
The only difficulty with this idea is that if you break everyday life down into a number of discrete tasks the performance of these many tasks becomes a memory/processing problem and is far from insurmountable in the longer term (due to Moore’s Law etc.). We just need a better computer and that item will definitely come with time.
Well, whatever the future holds, I am back involved with chess and enjoying it for its own sake. It is a real treat to play a game and win or lose be able to have GM strength engine go over the moves and check for blunders etc. The availability of this quality of analysis is definitely rapidly improving my game and I can’t help but think that in the future when we will have not only more powerful computers, but hopefully artificial intelligence to assist us, this process will be even more effective.
May that day be soon!
Dave Felton
