Author: Albert Silver
Date: 06:48:19 04/26/02
Go up one level in this thread
On April 26, 2002 at 09:38:12, Roy Eassa wrote:
>On April 26, 2002 at 09:02:28, Albert Silver wrote:
>
>>>I think 10-20 yrs is a long shot.
>>>If we take Fritz for example, we see that it releases new versions every yr,
>>>from the last version (DF) of this strong engine, Fritz7 is (arguably, but maybe
>>>not so)50+ avg stronger, it is almost more than 80-100 elo better than Fritz6
>>>(arguably again).
>>>
>>>We can perhaps state about the current version of Fritz7 that it is almost the
>>>same strength as lets say a human rated 2600-2650 (arguably). So with the
>>>current trend we can assume that when Fritz11 comes out at yr 2006, it will be
>>>(arguably) 200elo more than the current version which will be around 2800-2850
>>>(ahem arguably), and with much more faster processors, this could turn out to be
>>>even stronger.
>>>
>>>So maybe in another 4yr we can have Kasparov-kramnik running on our handheld PC.
>>>
>>>not too fictional IMO ;)
>>>
>>>pavs
>>
>>I don't know that I agree with your assessment Pavel. First there is the 80-100
>>Elo improvement you claim for Fritz 7 over Fritz 6. This increase is that touted
>>by the SSDF, but as has been generally acknowledged, the results of comp-comp
>>games aren't the same as comp-human games. The 80-100 Elo only represents it's
>>ability (performance) to beat other programs on other machines, but doesn't mean
>>that it will improve its performance over humans by the same margin. In fact I'm
>>sure it doesn't.
>>
>>Have you ever logged on ICC to see the games between some of the known computer
>>killers? One of them is rather interestingly the
>>recently-returned-from-retirement GM Mecking. I have had the opportunity to
>>observe several of his games and in particular noticed how he went about beating
>>them, when he did beat them. He holds his own in _blitz_ games against them mind
>>you, where they should theoretically be having a field day, and is undoubtedly
>>the key to the colossal rating he has on the server. He has also been about this
>>for a couple of years, and his win-loss ratio against them hasn't changed much
>>despite a few upgrades in software versions, and the much faster hardware. Also,
>>he always plays the highest rated accounts there. You should note that he is far
>>below 2700 FIDE and no longer at his best.
>>
>>On a side note, he has been visiting Rio consistently lately in order to have
>>intensive chess software training with yours truly so I expect to see his form
>>recover somewhat if he sticks to it. I'm also teaching him how to use the mouse
>>with ONE hand BTW.
>>
>> Albert
>
>
>Albert, that's _very_ interesting info on GM Mecking -- thank you for posting
>it!
>
>You have been quite silent on this board in recent weeks. I hope this signals
>your return in full force!
My silence was due to a general computer failure and inability to access the
Internet from home for a month or so. My desktop is still down though I have now
managed to hook up my laptop to my cable connection and my actual connection has
stabilized in the last week or so (Murphy's Law had it so that a left jab -
desktop dead - was deemed insufficient and followed up with a right uppercut -
broadband dead too). So yes, despite hating intensive typing on my laptop's
keyboard, I'm back online.
Albert
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.