22-Year-Old Norwegian Magnus Carlsen Is the New World Chess Champion 131
ardmhacha writes "Magnus Carlsen was able to force a draw in the 10th game of the World Chess Championship to claim the title with a 6.5 — 3.5 score (3 wins, 0 losses, 7 draws) over Viswanathan Anand. Carlsen became the youngest ever World No. 1 in 2010, but withdrew from the 2012 championship cycle and so has only now been able to add the World Champion title to his No. 1 ranking. He won three games and lost none. His first two victories came when he was able to convert small advantages in the endgame into wins. The third (in game 9) came after a blunder from Anand."
News for Nerds (Score:4, Insightful)
Re:How does he do against computers? (Score:5, Insightful)
Re:So what? (Score:1, Insightful)
Whether or not a better solution exists has nothing to do with whether or not the chosen solution is good. Nice non sequitur.
Re:WH Pushes Next Year's Enrollment Period Deadlin (Score:4, Insightful)
Maybe he's just earning a few bucks:
http://tech.slashdot.org/story/13/11/20/2218252/bp-hired-company-to-troll-users-who-left-critical-comments [slashdot.org]
Re:How does he do against computers? (Score:4, Insightful)
At the time, Deep Blue was the 259th most powerful supercomputer in the world with special purpose chess chips, a regular desktop today would be strong but not that ridiculously much stronger.
That is true, but software has also improved. We have better chess algorithms (especially pruning algorithms). But, even more importantly, we have better databases of previous games, and opening moves. Playing good chess has less to do with thinking, and more to do with remembering, than most people realize.
Re:How does he do against computers? (Score:4, Insightful)
1. Before the match, the computer (and computer programmers) analyzed all of the historical games by Kasparov and his most favored openings; any human at the level of Kasparov will have a very long footprint of history, while Kasparov didn't have any historical games of the computer to look at and to analyze
2. Both matches (1996 & 1997) ended after 6 games with the computer only winning by a 1-2 points, even without #1
3. "The rules provided for the developers to modify the program between games, an opportunity they said they used to shore up weaknesses in the computer's play that were revealed during the course of the match. Kasparov requested printouts of the machine's log files but IBM refused, although the company later published the logs on the Internet" (wikipedia Deep Blue page). I don't think this should have been allowed; the software should be true AI and learning without assistance
4. "Kasparov demanded a rematch, but IBM refused and dismantled Deep Blue." (wikipedia Deep Blue page) Kasparov and others never had another chance to beat it, after finally having a small history of games to analyze its playing style.
However, despite this, I think that a computer will most likely still reign supreme, but to be completely fair, I think it would require a history of games for the opponent to analyze and no human intervention during the match. However, the programmers can add in a "learning" module of some sort that analyzes each game afterwards, but no human intervention (e.g., programmers tweaking lines of code) is allowed during the match of games -- only before or after.
And on a related note, my main gripe with Watson was the physical responsiveness. There were times when the human hand reaction time could just not match the computer physically.
I would like to see a computer play blitz games against a world champion, as long as my gripe with Watson is ensured that they can't move physically faster than a human's reaction time.
Re:A champion may not even exist (Score:4, Insightful)
In this context, a total order satisfies transitivity. But being "better" in chess doesn't necessarily satisfy this property. What this means is that on average, player B wins against player A more frequently, and player C wins against player B more frequently, but player A could also win against C more frequently, making it impossible to state that any single player is the "best." This can occur because different players can exhibit particular strengths and weaknesses in different aspects of the game.
Note that it is important to talk about the above in terms of 'average' performance. Although chess is deterministic, there are random sources of variation in skill, in that a given player does not consistently choose the move that reflects their true skill level (i.e., they sometimes make a mistake, or they have a flash of insight).
For an interesting, rather counterintuitive, and simple example of non-transitivity, see http://en.wikipedia.org/wiki/Nontransitive_dice [wikipedia.org] .
Re:How does he do against computers? (Score:3, Insightful)
For some reason a computer playing from a database of pre-selected human games just doesn't sit as well for me as if the computer were actually finding the best moves through it's own calculation.
Also of note is that even with the massive database and relentless calculation, humans can beat computers at correspondence chess where the humans can spend enough time to calculate out everything just like the computer does. It's the time limit that makes their calculation so strong.
But to answer your question more directly, computers are rated somewhere around 3500 (although their rating has more to do with beating other computers), while Magnus Carlsen is rated 2870. However despite a 600 point rating difference, I'd expect he'd draw the majority of games against computers in a match.