On Wed, Jun 15, 2011 at 11:52 AM, Jacques Basaldúa <[email protected]>wrote:
> Don Dailey wrote: > > > only their own subjective "explanations" about why humans > > play much better at fast time controls. > > This is exactly the opposite of what I wrote: "that makes > humans *uneasy*." Of course, humans play better with slow > time settings. Blitz is very hard for humans. That's why > Zen is 5d at blitz and 4d at slow. Nobody is arguing that. > I wasn't directing that specific comment toward you, you did not give any subjective explanation for a particular belief. If you go back a few years you will see this topic was hotly debated - many were saying humans play stronger at fast time controls. The arguments presented were all just suppositions and I was alluding to that. Yes, you specifically stated that humans are uneasy at fast time controls and I could see that you understood that concept However you DID imply that the long time controls put the bot in some kind of zone where they will play especially good and I don't think I agree with that. > > but nobody has ever backed this up with anything > > concrete. > > Well the fact that bots scale is established, it is the > humans scaling which sounds reasonable but is not backed > up by evidence. > Bots scale very well, but the point is that humans scale much better. It's backed up by overwhelming empirical evidence from computer chess It's not about playing too fast to be "comfortable", it occurs at all levels. At speed chess humans are poor. At game in 1 hour computers are still better than at game in 2 hours, etc. I personally believe the reason is that humans make far better use of their time. We know when to use the time and when not to. We can play a move instantly without worrying that we missed something but we can also focus enormously well when we "suspect" there is an issue that needs investigation and we know what and where to look. Computers cannot yet do any of this except in the crudest ways. But that is just my theory on why, it could be nonsense. Just because I think it makes sense to me doesn't mean it isn't nonsense. Someone will no doubt present a plausible explanation of why computers play better than humans with more times but it will only be a theory and opinion just like the one I gave for the opposite. However what is not in question is that computers compete MUCH better as the time control gets faster. That's not a theory. > > And maybe the fact that bots scale is ignored by the > challenger whose knowledge about computer go I don't know. > I just pretended to warn him that he will face a much harder > opponent than he expects. And wish him luck. > > I think the bot will do well, but I would guess it will drop at least 1 dan in strength (relative to human) because of the long time control, as compared to the fast games. I just hope we are not too disappointed - but I'm cheering for the bot of course. Don > > Jacques. > > > > > _______________________________________________ > Computer-go mailing list > [email protected] > http://dvandva.org/cgi-bin/mailman/listinfo/computer-go >
_______________________________________________ Computer-go mailing list [email protected] http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
