Author: chandler yergin
Date: 01:39:39 02/06/06
http://www.aarontay.per.sg/Winboard/weaktablebase.html Quoting: Later you will see that in fact, in some positions, use of tablebases actually lead to weaker play, or may cause the engine to spend more time before finding the best move. Endgame tablebases don't take into account castling Another possibility that might occur results from the fact that the Nalimov tablebases used do not take into account castling. 3) Tablebases don't take into account the fifty move rule A much more serious problem in my view is that Chess engines don't take into account the fifty move rule. Notice though this is a problem because Nalimov endgame tablebases is a Distance to Mate (DTM) tablebase. Distance to Conversion (DTC) tablebase theoretically could solve this problem, just by checking that mates that end in the same tablebase do not exceed 50 moves.) However the Thompson tablebases current doesn't account for this. 4) "Knowing too much" problem This results mainly from positions when one side with tablebases resigns too early when it sees a loss in the tbs when the other side (human or program without tbs) might not be able to win anyway. 5) Tablebases slow down the search Probing the tablebases slows down the engine because it needs to access the hard-disk (despite tb hashes) this can be very costly depending on how much the engine probes in search1. This is most difficult to assess. We know that probing tablebases lowers the speed in nodes/seconds and usually the depth displayed in the principle variation. What does this mean? "Obviously (and logically) engines probing tablebases tend to play into won 5 piece positions (and avoid lost 5 piece positions), hence their high rating in such games. However, in the preceding endgame phase with still more pieces on the board tablebases somehow must be a sort of liability, so that engines using them lose games during this phase over proportion. Apparently statistically for each game they successfully maneuver into a won 5 piece position, they also lose another one during the preceding endgame, so that overall this effect more or less cancels out the advantages of tablebases."
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.