Folks, some news! I received an email a couple of days ago from Rex Gooch who has been working on word squares. While looking up some of our old discussions ...
Any of you guys have easy access to a multi-processor system such as a grid or a loosely coupled array? I have a tensquare program running that looks like...
Hi, As far as I know, there are in existence 5 different sets of Boggle cubes that were used over the years. Is this correct? I've attached my five files...
Using the OSPD and the small Boggle grid (1993 version), the number of words of lengths 5 and up that appeared in 100,000,000 random shakes of the cubes: 5...
Using the OSPD and the small Boggle grid (1993 version), the number of words of lengths 5 and up that appeared in 100,000,000 random shakes of the cubes: 5...
Is there any endgame solving software out there that is freely distributed and will run on a PC, or that is web-based? I'm looking for something at least as...
... You mean this one? http://bigftp.ocsystems.com/sim.html Apart from that one I don't know of any either. Have you asked GVC if he could fix whatever the...
... http://www.davidst.com/attic/savant.zip Savant goes into endgame solving mode when set to level 20, its highest playing level. Was this the first or were...
Steven, ... Savant was published as LanWords back in 1992. It was a crossword puzzle game for Windows that featured configurable rules (so you can configure...
... I had some trouble unpacking it. WinXP's built-in zip says it is corrupt, but unix unzip coped with it fine. You wrote this in C? Do you still have the...
... This is curious. I checked it and you're right-- XP says it's corrupt. Then I downloaded it to my desktop and tried opening it with XP and it worked....
... Maven's first endgame player was in 1986. By 1990 it had evolved to the point where it outplayed human champions at tournament time controls. (Figure 1 ...
sheppardco@...
Apr 12, 2004 6:20 pm
674
Hi David, I read your message about how important is skill versus luck in Scrabble. IMHO, the method you devised could not be very appropriate, since the kind...
Hi Enric, ... This is what I expected too. I hoped that this sort of cheating would be a conservative (overestimating) model of skill in a perfect player. You...
I, for one, am quite surprised at the 66% margin for a cheating program as David described. I would have thought the difference would be larger. Interesting to...
sheppardco@...
Apr 14, 2004 4:20 am
677
... That is interesting! Then again, we shouldn't give the cheater too much credit as it suffers dearly from the classic horizon effect. Anything significant...
Without rack heuristics to guide the use of the blanks and the Q and the J (not to mention all the other letters), I would be quite surprised if the "cheater"...
David, If the point of the exercise is to find an upper bound on programatic performance at Scrabble(R), then: 1. The cheat program really should have an...
Hi Steven, ... I agree for a simple-minded cheater such as the one I implemented. It was more work than I was willing to put into it at the time. A full...
... Maven has had this in research versions since 1996, and a commercial product was recently release that has this feature. It is the downloadable Scrabble ...
sheppardco@...
Apr 16, 2004 2:39 pm
683
... today. With an intelligent simulation controller, it is not necessary to pare down the move list much. Maven can admit 100 moves into its simulations and...
sheppardco@...
Apr 16, 2004 2:56 pm
684
... I'm not convinced a static evaluation can't be effective. While I think it's unlikely that somebody will create one as accurate as a bona fide simulation,...
Brian, Congratulations on the new version on real.com (yes, you said it was available on real.com) ;-). Is it correct that the static evaluation used to...
... It would be a lot of work. The source code was archived away more than ten years ago. ... Mr. Sheppard reports results of 60% to 63% over a static...
... Not really. The greatest gain is when you simulate *two* moves, and each move simulated beyond that point provides less benefit. Maven is already way...
sheppardco@...
Apr 17, 2004 3:37 am
688
Does Moore's Law provide any benefit in letting you simulate deeper (i.e., more plies) rather than broader (more different moves)? ... From: SheppardCo@......
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
