Author: Robert Hyatt
Date: 14:51:45 08/16/05
Go up one level in this thread
On August 16, 2005 at 17:44:59, Dann Corbit wrote: >On August 16, 2005 at 17:33:29, Robert Hyatt wrote: > >>On August 16, 2005 at 17:26:28, Uri Blass wrote: >> >>>On August 16, 2005 at 17:16:33, Paolo Casaschi wrote: >>> >>>>>I know that usually when program improve they improve in all time controls. >>>>>I do not know of evaluation changes or search changes that make programs weaker >>>>>at blitz but stronger at long time control. >>>>> >>>>>In thoery it can happen but I need to see a proof for it and I believe that >>>>>fabien mainly test in blitz time control(he can correct me if I am wrong) >>>>>because usually productive changes in blitz of adding knowledge to the >>>>>evaluation are also productive at long time control. >>>> >>>>Do you have any proof or evidence that there is some correlation between blitz >>>>strenght and slower speed strenght? >>>>If you dont, then we can only compare assumptions and I tend to agree with Bob >>>>Hyatt since the same non-correlation is evident with humans and because common >>>>sense... >>>> >>>>--Paolo >>> >>>Of course there is correlation. >>> >>>Look at every rating list at long time control and you can see Shredder,Fruit >>>Fritz,Junior,Hiarcs at the top of the list. >>> >>>Now look at rating list at blitz. >>>What do you see? >>> >>>Surprise for you >>>Again the same programs. >>> >>>You do not believe it? >>>Here are 2 rating lists one for blitz and another one for longer time control. >>> >>> >>>http://www.husvankempen.de/nunn/rangliste.html >>>http://www.miko42.de/turniere/blitzturniere/blitzrangliste.html >>> >>> >>>Uri >> >>Did you even look at your data? The programs do _not_ finish in the same order. > >On the other hand, there is definitely a strong correlation between strength at >standard time control and strength at blitz. I did not calculate it, but I >guess that it is fairly close to 1. it is cause and effect. If you are strong at blitz, you are strong at longer time controls, and vice-versa. But nothing says the two are tied together, as I have seen tons of cases where human A beats human B at blitz all the time, but never wins at longer time controls. And vice-versa. And when you factor in parallel search, it can get far worse... Yes a program ought to be able to "self-tune" to optimize performance for whatever time control is being played. In reality, this is a very _tough_ assignment since the hardware platforms vary so wildly, NUMA vs Non-NUMA, multiple-core vs single core, shared cache vs duplicate cache, different cache coherency protocols produce different overhead characteristics, 32 vs 64 bit, different sizes of TLBs, operating system capabilities or lack of them, etc. So while the program "ought" to be able to self-tune, in reality, I've never had the time or energy to try to develop such an algorithm. It is a project by itself, in fact...
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.