Author: Matt Frank
Date: 18:08:10 01/29/99
Go up one level in this thread
>What new algorithm are the current software vendors using that was not used in >Deep Blue? Micro assembly tweaks of existing algorithms mean *nothing* because >chess is an exponential process. I don't think it's the algoritms so much as the chess specific knowledge, endgame specific, opening book knowledge, anti-human strategy (open positions) etc. Oh by the way, Deep blue would basically have 5 minutes of processing time to be used for the whole game (consideration of it's moves and mine. The idea is that for the test to be appropriate to my suggestions then the pause button has to kick in, to prevent such massive permanent brain buildups on my machines' supposed time advantage. >>I think you don't know where the micros are now because the GMs are hiding. >>Anand said to a reporter after his match with Rebel 10, "No more, with >>computers." >He made himself a very bad deal when he agreed to speed chess (computer's >mightiest strength). Computers play very well even at 40/2 but I think he made >a choice such as that because of ego reasons. Look how Kasparov felt after >losing to a multi-million dollar mainframe. How do you think Anand felt after >officially losing to a micro -- fast time controls or not? My point being that you don't really have what you consider to be solid evidence on Micros because GMs are in a ducking mood. This brings me back to my original question. How do we get them to play. Please discuss. >No matter how far the current programs have advanced, from a purely mathematical >standpoint I can tell you that they are no match for deep blue. Deep blue could >brute force search farther than current hardware with alpha/beta in the same >time frame. In other words, it would be *impossible* to beat that machine with >existing hardware. Again the Deep Blue team would have no more than 5 minutes of processor time for the whole game, allocated the way they wanted to. This would provide a very fair test based on what I believe are benefits in speed more than offset by strong programming (among the software developers). Matt Frank
This page took 0.01 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.