Author: Robert Hyatt
Date: 20:34:42 10/16/00
Go up one level in this thread
On October 16, 2000 at 16:48:02, Chessfun wrote: >On October 16, 2000 at 15:38:56, Robert Hyatt wrote: > >>On October 16, 2000 at 14:05:52, Chessfun wrote: >> >>>On October 16, 2000 at 00:53:06, Robert Hyatt wrote: >>> >>>>My take: let's wait until the thing is released and see how it does. Without >>>>beta testers that exert a bit of influence over the program's time allocation >>>>and book choices. >>> >>> >>>Can you prove this statement please; >>> >>>There are no beta testers who exert any influence over book choice or >>>time allocation. The program runs automatically on the server, the book >>>is set. >>> >>>Sarah. >> >> >>I base it on the following. I have played multiple gambit tiger clones. They >>are reasonably predictable in their time usage. With a "couple" of beta >>testers, the thing will go into a "deep think" that is _far_ longer than the >>time one would expect for a move. IE it moves consistently at 30-50 seconds per >>move, then takes 10 minutes. In a position where it did _not_ fail low. I >>believe that the operator simply wants to give it a chance to find something >>that may (or may not) be there. >> >>It is my opinion, with no proof of course. But it is _very_ common with _all_ >>engines. ChessPartner makes it trivial for the operator to influence things. >>I can do it with xboard if I thought that I somehow might know more about when >>to spent more time than Crafty does... >> >>As far as proof, simply play a few fully automatic games at (say) 30 30 or >>whatever time control you like, and then check the times. See if you see any >>case where it takes more than 10x the normal time per move, when the score did >>_not_ drop _or_ rise during that search. If you find such cases, I will >>certainly retract my statement. But in watching so many games of late, it >>is obvious that something goes on "from time to time". IE I see most programs >>taking 2x-3x on fail lows. And sometimes for other reasons. But not 10x or >>longer. >> >>IE I would certainly like to do the same if that is the way games are to be >>played. When I say Crafty is "automatic" I mean _automatic". It does >>_everything_ by itself, completely. > > >You initially made two statements. >One on time allocation. The other on book choice. >The statement above explains why _you_ believe the time allocation theory, >it makes no reference about book which was also part of your original statement. > Book is pretty easy to follow. It _could_ be a result of "book learning" but I have no idea how Tiger does that. It _could_ be a result of the operator choosing the particular opening. That is _common_ for all programs on ICC. >Then as to the statement you make. >I am not aware how in the (CP5) interface it is possible to make the >program think for a longer period of time that it chooses to. > >Now naturally you have no experience with CP5 as there are only some >25 copies in use, but assuming that you refer to these practices happening >in previous CP interfaces and therefore assume it can happen in (CP5) please >explain how?. In CP 4 it was easy. Several documented (to ICC folks) how easy it was to cheat using the chesspartner interface. I won't spell out the details, but I would assume they are obvious to anyone that has used it a bit. And I also am assuming that 'features' were not taken out, which would be unusual. > >Since the pgn's do not have a time per move there is no way to check your >statement without a long and laborious task at ICC. That is assuming that it is >possible per the previous paragraph, that the operator can influence time. > >Sarah. You will have to simply watch some games. And then determine if you can find positions that occur where the program took _far_ longer in the game than it will if _you_ play over the game. manual computer operators are quite open about doing this. And I don't give it a lot of thought, other than to think that were I doing the same, it would be more "equal" than having a human help only on one side. Older chesspartner users simply set things up for the program to 'observe' in the background while the human played the game, to choose the opening. Then, once they reached the opening position they wanted, they would click the "play" mode on and let the program take over. I don't know whether this has continued in the current CP version, but it was definitely a "cheaters delight" when it first was made available. :( I didn't release my "custom interface" that I use to give online analysis during live games on ICC, because it would have made cheating just as easy as the CP interface originally did. Bruce didn't release his interface for the same reason. Of course we knew someone would do one, and they did... My original question still stands. If the program searches between 1-2 minutes per move (I would assume that in a 30 30 game you can find out what it would like to target for the time) and then goes well over 10 minutes in a position that is _not_ getting worse for it, then it looks like manual intervention. I have seen it happen often enough for manually operated programs that it is not hard to spot. Most programs are pretty regular in their time usage, although you _must_ watch the game to follow the pattern. ICC times are no good since they don't factor in pondering time. IE Crafty is _very_ solid in its time usage, until it finds an obvious move (easy move) or a fail-low condition. Some programs have a roughly 2x variance (they try to finish an iteration, or choose to not start another due to time left, etc). But I find over 10x unusual in a case where there is no 'problem' to be discovered... My obvious conclusion, based on lots of observation, is that the operator is exerting undue influence...
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.