[ List Latest Comments Only For Pages | Games | Rated Pages | Rated Games | Subjects of Discussion ]
Comments/Ratings for a Single Item
Captain Spalding Chess 
. Find an Elephant in your Pajamas.[All Comments] [Add Comment or Rating]btw, i use this zrf only to keep track of games played by email with other human players. even at maximum strength, maximum variety, and 3 minutes a move, zillions played unusually poorly. some of its moves defied explanation. the game itself is quite good tho, you definitely don't want to use zillions play as a basis for judging how successful a chess variant is.
The current release of the ZRF is now 1.2. By rewriting the neutral piece moving code to generate a single looping move entry instead of 64 move entries I have got this to the point where is loads and runs successfully on at least some Windows XP box (I would like reports from people running Windows XP about their experiances). This also has the side effect that the ZRF seems to run a bit faster, and examines more positions in the same time as a result. There turns out not to be an elephant problem as reported earlier -- that was a misunderstanding.
Mike Nelson wrote on Sun, Aug 18, 2002 03:54 PM UTC:I think this is a Zillions issue. I ran a Zillions vs. itself game on my machine (A four-year-old Celeron processor running Windows 98 SE) using these settings: Expert stength, Moderate variety, 30 seconds think time. While Zillions doesn't play nearly as well as it plays FIDE chess, for example, there were some good moves. Black set up a combinatation to fork White's Box and Pajama with the Great Elephant, then won by exchanging an Elephant several times, able to get a replacement while White could not. (Unfortunately, I didn't save the .zsg file.) I'm not yet sure how to rate this game, but I think this may well be the wierdest chess variant that still IS a chess variant.
While I have not made much study of the play of this game yet, it is the sort of game where Zillions (a), needs a lot of thinking time for, and (b), doesn't play particularly well, anyway. The Zillions board evaluation algorithm seems to be primarily material-based, and that doesn't seem to yield good results for this game.
M. Howe wrote on Sun, Aug 18, 2002 03:24 PM UTC:The Captain Spalding ZRF runs fine for me (Win 98se) but to those who can't run it, I'm not sure you're missing much. Peter has certainly done a great job writing the ZRF, and I saw no bugs or rules anomalies, but when I told Zillions to play itself, the resulting game was tedious, repetitive, and lackluster with nothing you could call a 'combination.' The boxes ran past each other toward the promotion zone, and proceeded to drop many bats and promote them to many BOOHs. The great elephants were exchanged early, and the pajamas didn't last long either. Finally, after 150 dull moves, black won. I know that Ralph Betza has invented many seminal chess variants and I respect him greatly, but unless what I saw was just Zillions completely unable to cope with whatever the strategy of this game must be, this doesn't strike me as a successful variant.
Peter - I just downloaded the ZRF 1.1 dated 8/15, and Black was able to pull three Elephants from his Pajama before I called it quits. I had no trouble running it with Zillions 1.3.1 and Windows 95. I saved a ZSG file if you need it. I did not test for the same situation for the Great Pajama or the Box. The graphics are a big improvement over the alpha version I got a while back.
The following from Jeff Mallett:
<p>
<i><blockquote>
Bad news -- it crashes for me on XP in 1.3.1 (and also crashes when I
tried 1.3). It doesn't crash when I use Zillions 2, which is in
development, so it's a problem already fixed.
<p>
I'm betting the difference is that Zillions 2 has had a lot of its
reading code rewritten to support parsing larger files. The XP
difference is just that some machines need to reallocate memory at
some point in the parsing and others don't. It's this reallocation
that was causing problems in Zillions 1.3.1.
<p>
Your ZRF itself is pretty small and elegant (23K), but when its 28
macros are expanded it is over a megabyte and contains about 175,000
symbols. Is there any way to simplify it a bit?
</blockquote></i>
<p>
I will look at simplifying it, but I have my doubts -- the thing was hard enough to write in the first place.
OK, it simply flatly blows up in Windows XP, then I have something I can tell the Zillions people. I hope they have an XP box in-house!
Joseph DiMuro wrote on Fri, Aug 16, 2002 11:12 PM UTC:I have the exact same problem; I also have version 1.31p, running on Windows XP. It's the same kind of error message that sometimes pops up when I'm running Internet Explorer (when I access certain websites). I have no idea what's wrong...
Version 1.31p, running under Windows XP. :-P
Revision of the ZRF is now 1.1, and the problem with Dust Demon capture and ko is now fixed.
Glenn, what's your rev of Zillions, and what operating system are you
running on? I had problems with the ZRF earlier, but they were fixed by
not combining self-capture and add-partial in the same move.
<p>
As for not repeating the board position, I'm only checking for the simple
case of undoing a previous neutral move -- Zillions isn't capable of
anything else, unfortunately (but then, human players are pretty limited
there too). Actually, I just noticed a slight error in the neutral
move code that if your opponent captures with a Dust Demon, it won't let
you move the Dust Demon back to its previous square right away. That I
can fix.
My Zillions gags on the zrf; I don't know why. It encounters a Windows problem and needs to close before it ever opens. First time I've had that issue when opening a zrf. BTW, I was working on a Captain Spalding when I saw your post. How did you solve the 'can't repeat a position' rule?
13 comments displayed
Permalink to the exact comments currently displayed.