Psion 3D Chess vs. Sargon IV on the Macintosh
r...@mva.cs.liv.ac.uk
Sargon IV Chess vs. Psion 3D Chess on the Macintosh
A Review by Richard K. Lloyd
At long last, my copy of Sargon IV Chess came through from MacSoft (a usually
very rapid and reputable Mac dealer), so I've decided to produce this long
document on the two best chess programs current available on any microcomputer
(no, I haven't seen Zarkov or AI Chess on the IBM PC yet - anybody willing to
supply me with review copies ? :-) ).
1. Review Machine
-----------------
A humble standard Macintosh Plus with 1MB RAM, an external 60MB hard drive
and two 800K floppies. No, I don't have access to a Mac II - at Apple UK's
current crazy pricing strategy, is that surprising ?
2. Configuration of Switcher
----------------------------
Being a die-hard fan of Switcher (MultiFinder is useless on a Mac Plus), I
set about configuring Switcher to run BOTH Psion 3D Chess and Sargon IV Chess
simultaneously. I configured both to use 256K each and they seemed happy with
that, but WATCH OUT ! Install Psion 3D Chess FIRST and Sargon IV Chess after-
wards otherwise you'll get a totally inexplicable System crash.
3a. Running Psion 3D Chess
--------------------------
Psion 3D Chess comes up with a 3-D board, Sound On, Analysis Off and the time
limit set to 6 seconds. This is exactly the opposite of how I like it and there
is no way to save a configuration file to change your startup preferences.
The 2-D pieces leave a lot to be desired, being faceless rounded objects for
the most part. This is a big shame, because the 3-D pieces are actually very
good indeed.
3b. Running Sargon IV Chess
---------------------------
Sargon IV Chess now does something when starting up that Sargon 3 never used to
do: display a pointless picture of an unshaven knight with the words
"Sargon IV" written on his helmet and then proceeds to very slowly (on my Mac
Plus I hasten to add) scroll the picture downwards off the screen. Now this
would be fine IF you could click on the mouse or press a key to skip this,
BUT YOU CAN'T ! It causes me to get angry with Don and Kathe for such an
irritating setup.
Some more disk access occurs (with an unfriendly blank screen and menu bar)
and you finally go into, yes you've guessed it, 3-D board mode. At this point,
the casual player would say "Ugh ! Those pieces are disgusting !" and stick
with Psion. The standard 3-D pieces are poor in my opinion, but you do have
the option of changing that (see later). Again, there's no way to save your
preferences - a major oversight, because Sargon IV was written in 1988 and
can't use Psion's excuse of being an "old-style" Mac program.
4. Playing a game
-----------------
The first thing to note is that both programs start with the Book switched on,
but only Sargon IV of the two can switch it off. If you don't do this, then
the first 9 or 10 moves against each other are done at lightning speed !
A major disappointment with Sargon IV's Move Analysis was quickly discovered:
because, like Sargon III, it is in a window, it is NOT shown in the 3-D board
mode. It should have reverted to the Psion style in that mode i.e. along the
front edge of the 3-D board instead.
Settling down with Switcher and the two chess programs also revealed a further
Sargon flaw (inherent in Sargon III too) - it is pretty hostile about recog-
nising the mouse click on the double-headed menu bar arrow to switch tasks.
This often also results in the pull-down menus NOT appearing, no matter how
many times you click on the menu bar. Don and Kathe really ought to have tested
this - I'm sure there are many Mac Plus users who still use Switcher.
During gameplay, I much prefer Sargon IV's (and III's) move analysis, but this
is more than outweighed by the highly irritating and UNREMOVABLE beep that
Sargon IV insists on making every time a piece is moved. Yes, you can zero the
volume on the Control Panel, but why should that be up to me ? Other than that,
the game options are very similar indeed. Both programs can take moves back
right to the start and both can replay moves from the start onwards. Saving of
the game (unfortunately not the configuration too) seems to be OK, but Psion is
a bit awkward reloading it - it has to be in the same directory as the main
program, it doesn't use a proper icon and quite often screws up after loading
and won't let you continue in a sensible manner.
Both programs have flexible level settings and can think on opponent's time,
though Psion has better spaced time intervals (I would have preferred to choose
the minutes/seconds per move myself in a dialog box) and supplies the useful
'Equality' mode. Neither program allows you to enter moves from the keyboard
(e.g. E4E5) - a poor omission. Sargon IV employs the click-drag-release
movement system (my preferred method) whereas Psion uses the click-drag-click
method, which is more cumbersome but rules out placement error I suppose.
Playing Strength
----------------
Up until I got Sargon IV, Psion 3D Chess had demolished all the computer
opponents I put against it. This included Colossus 4 Chess on a BBC Micro,
Gnu Chess running on the Acorn Archimedes and a fast UN*X mainframe,
Sargon III on the Mac, Chessmaster 2000 on an Amiga and, for a laugh,
Psion 3D Chess on a PC clone. I was never too strict on enforcing tournament
conditions in the tests, but I found somewhere between 15 and 30 seconds a
move about as much as I could tolerate without falling asleep.
There is a somewhat bizarre statement in the Sargon IV manual that says:
"[Displaying] The Search Window will slow down SARGON IV by about fifty
percent...". I can believe this if the 'Display Tree' option is checked,
but otherwise I think this is a completely erroneous assumption. The search
window is only updated when a move at the current ply has been considered
(involving either only the change of one number on the screen or a refresh
of the best line found so far) - this can be a gap of several seconds at the
higher ply levels.
I am very sad to report that Sargon IV appears to play no better than its
predecessor Sargon III. This seems almost outrageous in light of the claim on
the box: "Sargon's chess algorithm is the most powerful available on a
microcomputer" and a statement is made that Chessmaster (2000?) is beaten 85%
of the time, which isn't exactly difficult. I remain convinced that the best
chess program I've seen on a micro, on pure playing ability alone, is Psion 3D
Chess and I challenge the Spracklens to prove otherwise under strict tournament
conditions...
Although I haven't posted any games to prove my point, Sargon IV was beaten ALL
of the time by Gnu Chess on the Archimedes at the 5 seconds level and only
managed to overhaul Gnu Chess at about the 30 seconds point. This is because
Sargon IV is notoriously weak at low time levels - it doesn't manage to get
past the third ply most of the time. Psion 3D Chess was only beaten once and
drew once in 10 games at varying times (5, 15 and 30 seconds a move).
4. Features Checklist
---------------------
As a quick reference guide, here's a comparative checklist of the two programs:
Feature Psion 3D Sargon IV Comment
-------
Switcher compatible Yes Yes/No Sargon often ignores <->
Help Yes No Not really needed
Multi-language Yes No Good for foreigners !
3-D Board Yes Yes Psion's pieces better
Custom 3-D Pieces No Yes Sargon has 7 sets & editor
2-D Board Yes Yes Sargon's pieces MUCH better
Custom 2-D Pieces No No I wonder why Sargon can't ?
Opening Book Yes Yes Sargon's bigger and removable
'Novice' Level Yes Yes Is there any point to this ?
Timed Levels Yes (11) Yes (8) Big time jumps in Sargon
'Equality' Level Yes No Bad omission from Sargon
'Infinite' Level Yes Yes Great for problem solving
Fixed-Ply Levels No Yes Any point to this either ?
Mate In.. Solver Yes (8) Yes (15!) Nice to see in Sargon at last
Load/Save Game Yes Yes DOESN'T SAVE CONFIGURATION
Move analysis Yes Yes Sargon better, but NOT IN 3-D
Next Best Move Yes No Ultra useful for variatioms
Hint Yes Yes I never use this personally
On-screen clocks Yes Yes New to Sargon IV
Countdown clock No Yes Excellent for tournaments
Set up board Yes Yes Both good
Classic Games Yes Yes Most moves missed by both !
Problem Positions No Yes Handy for problemists
Copy Protection Yes No Psion hard disk-runnable
with know-how
Conclusion
----------
If you already have either Sargon III (and don't want a 3-D board) or Psion
3D Chess, then I wouldn't bother buying this program because it just doesn't
play strong enough chess considering it's the fourth major revision over the
last 10 years. However, Mac II owners might want to consider it for the colour
3D board (although the best picture on the box is of the IBM PC version !) and
possible (?) Multifinder compatibility.
In the UK, this game costs about 47 pounds which is clear overcharging on
behalf of the distributors and I don't feel I got my money's worth.
P.S. You might like to know that I copied (using a combination of a screendump
of Psion's 3D board, the Scrapbook, SuperPaint, Piece Designer and
ResEdit) Psion's 3D pieces on top of the standard 3-D set that comes up
when Sargon IV is first run. The board shading needed changing too, but
I now have virtually identical board/pieces in both programs !
Richard K. Lloyd, **** This is a MicroVAX II running VAX/VMS V5.1 ****
Computer Science Dept., * JANET : R...@UK.AC.LIV.CS.MVA or *
Liverpool University, * R...@000010500211.FTP.MAIL *
Merseyside, England, * Internet : RKL%mva.cs.l...@cunyvm.cuny.edu *
Great Britain. ****************************************************
Ronald J. Rangel
A comment about Psion Chess: I have it for my Atari 1040ST and it is
easily the strongest program I've played
on that machine. A good buy at $27.00.
Patrick L Faith
Sargon has some features for kids that are not present in the other packages.
For example in sargon, they enjoy the chess pieces with the big noses and
generally funny faces. Now this might not seem important, but to keep little
kids interest up, it helps a lot. Four year olds have no problem with the
user interface, which is a big, big block for them to get over. Any way, for
teaching kids chess, I think Sargon is excellent, in that it does not
distract them with to much silliness, like battle chess, etc.., while still
giving them some fun. If any one else has some comments about teaching kids
chess with computers, send it to me, I need the help.
P L Faith
Mark Lord
I'm looking for an easy to use chess program, and thus far have tried
(purchased) ChessMaster2000, which has very annoying copy protection,
making hard drive installation difficult (though not impossible).
I also bought ChessMaster2100, which claims to have NO COPY PROTECTION on
the package, though the diskette DOES indeed include a copy protection
mechanism (pretty slimey!). In addition, it does allow hard drive
installation, but has a STUPID screen that prompts me for some trivia from
the manual each time a fire it up. Enough to strongly discourage me from
playing impromptu games with it (my desk is too cluttered to keep the manual
around, and the stupid thing doesn't work inside DESQVIEW either!).
I have been contemplating getting SARGON IV, but the package does not
indicate whether or not it also has such nasties..
Would the original poster(s) be so kind as to let us all know ??
Thanks!
-Mark
--
+-------------------------------------+----------------------------+
| Mark S. Lord | Hey, It's only MY opinion. |
| utgpu!bnr-vpa!bnr-fos!mlord%bmers58 | Feel free to have your own.|
+-------------------------------------+----------------------------+
R. Scott V. Paterson
>I'm looking for an easy to use chess program...
I'm looking for a chess program that is animated. Not just that the chess
pieces move but that they are made in the form of their repective
pieces, ie a king is a king, and that they do animated battle. I've
seen this for the Amiga and it was called Battle Chess. Anyone know
if this exists for the mac or whether anyone is planning on making
it exist for the mac?
Thanks,
-rsvp
Patrick L Faith
protection, if you copy from the original disk to a hard disk. It does not
ask questions from the manual to verify users.
P L Faith
Dennis J. Kosterman
|
|(purchased) ChessMaster2000, which has very annoying copy protection,
|making hard drive installation difficult (though not impossible).
|
|I also bought ChessMaster2100, which claims to have NO COPY PROTECTION on
|the package, though the diskette DOES indeed include a copy protection
|mechanism (pretty slimey!). In addition, it does allow hard drive
|installation, but has a STUPID screen that prompts me for some trivia from
|the manual each time a fire it up. Enough to strongly discourage me from
|playing impromptu games with it (my desk is too cluttered to keep the manual
|around, and the stupid thing doesn't work inside DESQVIEW either!).
I (and my roommate) have had trouble with both "Chessmaster" programs.
I bought "Chessmaster 2000" when it first came out, and immediately dis-
covered two things about it that I didn't like: (a) the box lists a whole
bunch of features, with a disclaimer stating that certain features are
available only in certain versions (i.e., which features you get depends
on which computer you have). But they never make it clear, either on the
box or in the manual, which features your particular version does and
doesn't have -- you have to find out by trial-and-error. This remains
a problem with "Chessmaster 2100".
The second problem is more serious: at least on the Apple II version
(which is what I have), the program itself contains a very annoying bug:
it quite often makes the "move" A8*A8 (the Queen's Rook blinks, but remains
on the board). I find this intolerable -- it immediately invalidates the
game, and you have to start up a new one, in which the same thing may happen
again. It seems to me that it ought to be trivial to write a program that
(at least!) always makes legal moves! I can't believe that a program this
popular has such a basic bug.
Anyway, I soon lost patience and went back to "Sargon III" -- it may
or may not play better chess than "Chessmaster", but at least it always
makes legal moves, and allows me to do the same. In hundreds of games
against Sargon, I've never once had it do anything illegal.
My roommate recently bought "Chessmaster 2100" (again it's the Apple
II version), and it too has problems:
1. It has the same problem as "...2000" with documentation of the
differences between versions.
2. It doesn't appear to have the "A8*A8" bug, but it did, at least
once, refuse to allow him to make a legal move.
3. He was unable to make a backup copy, even though the box claims
that there is no copy protection.
4. There is no manual! Either it inadvertantly got left out of the
box, or there's not supposed to be one. The program is menu-driven, and
we were able to figure out the essentials without a manual, but it seems
as if there ought to be one. Also, using the menus is slow and becomes
very annoying after a while. There ought to be a way to bypass them, but
without a manual, we may never find it (if it even exists). If you *must*
use the menus, that's another black mark against the program in my book.
Conclusion: I am singularly unimpressed by both "Chessmaster" pro-
grams, and by Software Toolworks in general. Note: I make no claims as
to Chessmaster's *chess-playing* ability -- I'm not a good enough player
to evaluate that -- but as far as the basic things (the things you ought
to be able to take for granted), it fails miserably. When I want to play
computer chess, I'll continue to reach for "Sargon", not "Chessmaster".
Speaking of "Sargon", I've seen "Sargon IV" for the Macintosh, but
not for the Apple II. Is there an Apple II version? And if not, will
there ever be?
Dennis J. Kosterman
uwvax!astroatc!stubbs
Robert Fenske Jr
>|I'm looking for an easy to use chess program, and thus far have tried
>|(purchased) ChessMaster2000, which has very annoying copy protection,
>|making hard drive installation difficult (though not impossible).
know I can't install it on my hard disk, which is VERY annoying.
If anyone has done this with the Amiga version I'd like to know.
>|
> [stuff about ChessMaster 2100 & 2000 deleted]
>
> Conclusion: I am singularly unimpressed by both "Chessmaster" pro-
>grams, and by Software Toolworks in general. Note: I make no claims as
>to Chessmaster's *chess-playing* ability -- I'm not a good enough player
>to evaluate that -- but as far as the basic things (the things you ought
>to be able to take for granted), it fails miserably. When I want to play
>computer chess, I'll continue to reach for "Sargon", not "Chessmaster".
>
environment; I wish some things were organized better, but
that's personal preference. I haven't had any illegal moves
from mine. In general, I've been satisified with it. It plays
pretty good chess also. I've never played a Sargon III version
of any sort, so I can't compare the chess-playing ability. But
I agree that the user interface (and manuals, etc) is very
important, and I wish no program any annoying copy protection
scheme whatsoever.
--
Robert Fenske, Jr. Sw | The Taming the C*sm*s series:
Electromagnetics Division /R---\ |
Southwest Research Institute | I | | "The Martian canals were the
dfsun1.electro.swri.edu 129.162.160.4 \----/ | Martian's last ditch effort."
Bruce Lowerre
> In article <29...@astroatc.UUCP> stu...@astroatc.UUCP (Dennis J. Kosterman) writes:
> >|I'm looking for an easy to use chess program, and thus far have tried
> >|(purchased) ChessMaster2000, which has very annoying copy protection,
> >|making hard drive installation difficult (though not impossible).
> I have the Amiga version of ChessMaster 2000 and as far as I
> know I can't install it on my hard disk, which is VERY annoying.
> If anyone has done this with the Amiga version I'd like to know.
> >|
> > [stuff about ChessMaster 2100 & 2000 deleted]
> >
> > Conclusion: I am singularly unimpressed by both "Chessmaster" pro-
> >grams, and by Software Toolworks in general. Note: I make no claims as
> >to Chessmaster's *chess-playing* ability -- I'm not a good enough player
> >to evaluate that -- but as far as the basic things (the things you ought
> >to be able to take for granted), it fails miserably. When I want to play
> >computer chess, I'll continue to reach for "Sargon", not "Chessmaster".
> >
> The program I have is integrated fairly well into the Amiga
> environment; I wish some things were organized better, but
> that's personal preference. I haven't had any illegal moves
> from mine. In general, I've been satisified with it. It plays
> pretty good chess also. I've never played a Sargon III version
> of any sort, so I can't compare the chess-playing ability. But
> I agree that the user interface (and manuals, etc) is very
> important, and I wish no program any annoying copy protection
> scheme whatsoever.
I've had nearly the same experiences with Chessmaster 2000 and Sargon IV
on the Mac. The inablility to run Chessmaster 2000 from a hard disk is
very annoying. Also, I once had the program in an obviously drawn
position. It always refused my draw offer so I decided to play it out
until we reached a three time repeated position. After I made the move
which repeated the position three times, it went berserk and made an
off-the-wall illegal move. I finally got so fed-up with it, I sent the
floppy back to Toolworks, thinking it had bugs, asking for an explanation.
BIG MISTAKE. I never got a reply from them nor did I ever get my floppy
back. DON'T BUY ANYTHING FROM TOOLWORKS! They're a rip-off.
As for Sargon IV, GREAT!! The only thing, as has been previously posted,
is that one cannot set the initial conditions. The unshaven face always
comes up at initialization; I think it's entertainment while the program
initializes. And, the 3D board always appears. I prefer the 2D board.
Stuart Cracraft
the two "standard" PC chess programs Sargon 4 and Chessmaster 2100.
On an IBM PC XT these programs are, at most USCF 1680 and 1800
respectively -- which is to say, rather poor.
To those of us who know what can be done on the PC, Sargon 4 and
Chessmaster 2100 discussion is distressing.
There are at least FOUR programs significantly stronger than Sargon
4 and Chessmaster 2100:
AI Chess
Zarkov
Psion
Rex Chess
These programs vary from 100 to 320 points stronger than Sargon 4 and
Chessmaster 2100, running on THE SAME MACHINE. Of these, AI Chess is
by far the strongest. At Software Toolworks it defeated a former US champion,
while running on a 25mhz 80386, and defeated a computer chess specialist
master who beat Belle in a much-publicized tournament-encounter, and
drew a 2359 master. [ A PC DID THIS !! ]
If you have Sargon 4 or Chessmaster 2100 and are dissatisfied with them,
you should definitely find out more about the above programs.
Why settle for mediocrity in PC chess when better is out there?
--Stuart
Kevin Clendenien
>In article <14...@dfsun1.electro.swri.edu>, fen...@dfsun1.electro.swri.edu (Robert Fenske Jr) writes:
>> In article <29...@astroatc.UUCP> stu...@astroatc.UUCP (Dennis J. Kosterman) writes:
>> > Conclusion: I am singularly unimpressed by both "Chessmaster" pro-
>> >grams, and by Software Toolworks in general. Note: I make no claims as
>
>BIG MISTAKE. I never got a reply from them nor did I ever get my floppy
>back. DON'T BUY ANYTHING FROM TOOLWORKS! They're a rip-off.
>
2000. I have Chessmaster 2000 for the Atari ST, and I'm very happy with
it. Mind you, I'm not refering to it's playing ability, since I'm hardly
qualified to comment. What I am refering to is all of the nice features
that go with it. You can save every initial condition: 2-D or 3-D display,
view of board, playing strength, time clocks on or off, and much more...
You can configure, and save, the color of the board, and pieces. You
can save any game in progress. It comes with 100 past games on disk. These
include Grandmaster games, as well as some computer vs computer games.
You can have the program give its analysis of the current game, or games
loaded off of disk. In short, it is a very well thought out, and put
together game... At least the Atari ST version is. Yes, the ST version
is copy protected, but that is the only thing about the program that I don't
like.
--------------------------------------------------------------------------
kcle...@silver.bacs.indiana.edu Kevin Clendenien
BLoomington Atari ST users group BLAST, President
BLAST BBS - (812) 332-0573 FNET node #141
"Of course any opinions or views stated above do not necessarily represent
the official position of any person, or organization other that of Kevin
Clendenien."
--------------------------------------------------------------------------
Bruce Lowerre
> In article <36...@apple.Apple.COM> low...@Apple.COM (Bruce Lowerre) writes:
> >In article <14...@dfsun1.electro.swri.edu>, fen...@dfsun1.electro.swri.edu (Robert Fenske Jr) writes:
> >> In article <29...@astroatc.UUCP> stu...@astroatc.UUCP (Dennis J. Kosterman) writes:
> >> > [stuff deleted]
> >> > Conclusion: I am singularly unimpressed by both "Chessmaster" pro-
> >> >grams, and by Software Toolworks in general. Note: I make no claims as
> >
> > [stuff deleted]
> >BIG MISTAKE. I never got a reply from them nor did I ever get my floppy
> >back. DON'T BUY ANYTHING FROM TOOLWORKS! They're a rip-off.
> >
> I'm sorry to hear that you have had such bad experiences with Chessmaster
> 2000. I have Chessmaster 2000 for the Atari ST, and I'm very happy with
> it. Mind you, I'm not refering to it's playing ability, since I'm hardly
> qualified to comment.
Chessmaster 2000 on the Atari ST may be OK. I don't know since my only
experience has been with the Mac version. I got frustrated with their
software and am pissed off at them as a company for not even answering
my inquiry and stealing my floppy. There are guidelines to follow when
developing software for the Mac. Toolworks violated many of them to the
point that the Chessmaster 2000 on the Mac is nearly unusable.
Louis Blair
and asker of "deep and significant" questions) writes:
>On an IBM PC XT [Sargon 4 and Chessmaster 2100]
>are, at most USCF 1680 and 1800 respectively
>AiChess is 300 points higher than Sargon 4, 200 points
>higher than Chessmaster 2100, and about 125 points
>above Zarkov and Psion.
Louis Blair (the well known "chess groupie" who writes
"mass verbiage" and has never written any portion of a
chess program) asks the following not-very-deep questions:
(1) Who did the testing that produced these conclusions?
(2) Did this person once tell us, "The ... score [of a match
recently contested between Fidelity Mach III ... and ZARKOV,
a new chess program available for the IBM PC] was 7 points
for the master machine and 3 points for ZARKOV. This works
out to about 2153-2190 rating points for ZARKOV when running
on this configuration"? What do you think about the correctness
of this calculation?
(3) Did the test conductor have any financial involvement
in any of the tested programs?
(4) Do you know who pays for the advertisement in Chess Life
that reads, "Zarkov, expert strength chess, IBM PC and
compatibles. $25-$35, depending on configuration. MSM,
5 Via Amistosa ..."? Do you know how long that advertisement
has been apppearing in Chess Life? Do you know how much,
it has cost to run that advertisement?
(5) Was the test conductor involved in the writing of the opening
book for any of the tested machines? Had any of the other
programs been used as opponents while the opening book for one
of the programs was being written?
(6) Is it possible to improve a machine's performance against
specific other programs by making changes in the opening book?
(7) How many games were used in the tests? What accuracy
could one expect from rating calculations based on this number
of games?
(8) What time control was used for the tests? Were the tests
between human and machine or between machine and machine?
Are machine-machine tests at one time control an accurate way
to predict what results one could expect between a machine and
a human at a significantly different time control?
Stuart Cracraft writes:
>[Zarkov version] 1.30 is now in release
(9) Do you remember this note?
"Subject: NACCC final crosstable
Date: 17 Nov 89 03:42:28 GMT
1 Hitech 2400 2513 9+b 6+w 2-w 4+b 5+b 4.0
2 Deep Thought 2551 2508 7+b 8+w 1+b 5+w 3-b 4.0
3 Mephisto X 0 2309 10=w 5-b 6+b 7=w 2+w 3.0
4 Bebe 2150 2169 8-b 9+w 7+b 1-w 10+b 3.0
5 Rebel X 0 2307 6=b 3+w 8+b 2-b 1-w 2.5
6 Cray Blitz 2250 2234 5=w 1-b 3-w 9+b 7+w 2.5
7 Phoenix 0 2053 2-w 10+b 4-w 3=b 6-b 1.5
8 BP 0 2014 4+w 2-b 5-w 10=b 9-w 1.5
9 Novag X 2092 1999 1-w 4-b 10=w 6-w 8+b 1.5
10 Zarkov X 0 1959 3=b 7-w 9=b 8=w 4-w 1.5"
Which version of Zarkov was this? Is 1959 "expert strength"?
Stuart Cracraft writes:
>DT's surprising 29. Qg5 against Cooke at the Software
>Toolworks is apparently not too difficult to find, at least
>for a machine. The PC program "ZARKOV" found and
>played 29. Qg5 in only 5 ply of search (not counting
>extensions for captures and checks). The winning Qg5
>was found at the second to last variation to be searched
>at the 5th ply, in 4 minutes 30 seconds, on a "nothing great"
>16mhz 80386.
(10) Why stop with Zarkov? Why not see how low we can
go? (I tried it on the approximately five-year-old Novag
Super Constellation at level five (approximately two minutes
per move). It played 29 Qg5 after almost exactly two minutes.)
Does anyone have Sargon III or one of those cheap Radio
Shack machines?
(11) How much agreement is there between the ratings
posted by Stuart Cracraft and the opinions of Larry
Kaufman, primary author of Computer Chess Reports?
The latest issue will soon be available.
Dennis J. Kosterman
|
|There has been so much fanfare and hoopla in this newsgroup about
|the two "standard" PC chess programs Sargon 4 and Chessmaster 2100.
|
|On an IBM PC XT these programs are, at most USCF 1680 and 1800
|respectively -- which is to say, rather poor.
|
|4 and Chessmaster 2100:
|
| AI Chess
| Zarkov
| Psion
| Rex Chess
|
|
|Why settle for mediocrity in PC chess when better is out there?
Well, I can think of a few reasons:
1. Some of us (like me) are mediocre chess players. My best USCF
rating (when I had one) was 1328, and it'll be years before I reach 1800,
if I ever do. Sargon and Chessmaster are plenty strong enough for me -- I
still have trouble beating Sargon on the middle levels.
2. Some of us (like me) do not have an IBM PC XT, or a Macintosh, or
any of the fancy (and expensive) new computers. All I've got is a lowly
Apple IIe, and my budget tells me that's all I'm going to have for quite a
while. I'm not aware that any of the programs you mentioned are available
for the IIe.
3. Even supposing I had an IBM PC XT, how widely available (and inex-
pensive) are the above programs? You can walk into any store that sells
software and buy Sargon or Chessmaster for <$50. I've never seen any of
the 4 you mention (although I've heard of them).
Dennis J. Kosterman
uwvax!astroatc!stubs
Ronald J. Rangel
Chessmaster away. I would recommend playing it first before forking
over the $$$.
Ron
William Tsun-Yuk Hsu
>|
>|Why settle for mediocrity in PC chess when better is out there?
>
> Well, I can think of a few reasons:
>
> 1. Some of us (like me) are mediocre chess players. My best USCF
>rating (when I had one) was 1328, and it'll be years before I reach 1800,
>if I ever do. Sargon and Chessmaster are plenty strong enough for me
Ummm, I'm not sure I agree with this attitude. I've never risen far
above mediocrity myself (1700+ when I stopped playing 4 years ago,
higher-rated chessplayers in their more charitable moods considered
me dangerous but hopelessly inconsistent :-)). My attitude had always
been that I'd rather lose a good game that I could learn from and
show people than bungle my way to a win. I couldn't consistently
beat my old chess computer at regular time controls, but I still
felt severely limited by it. I used it more for study and analysis,
and I was frustrated by the small (and relatively mainstream) opening
book, and its inability to find a good plan in many positions.
Bill
Owner of Many System Processes
garbage. I don't know if I got a bad copy, but:
From: mrhar...@trillium.waterloo.edu (Mike Harrison)
Path: trillium!mrharrison
1) It makes many illegal moves in 3-D mode (I've never seen it do illegal
moves in 2-D), such as moving a pawn four square forward and four
squares left in one move!
2) It often switches sides without telling me. I'll be playing, and next thing
I know, it won't let me move my white pieces, only Black's, and it will
play white.
3) When a checkmate occurs, if you try to move a piece, the game crashes. This
should not happen in a well-written program. Users are not supposed to be
able to crash computer programs by using it normally.
4) It sometimes will not let me make legal moves. I once tried to queen a pawn,
and it kept moving my piece back and saying "Illegal move" in it's smug,
annoying computer voice it has. (I especially hate the way it says "
You lose". MAkes it sound like it thinks it's the greatest, or something)
5) In the directions, it says there is a picture file called "template.pic"
on the games diskette, for making custom chess pieces. This file does
not exist on either disk.
6) It says it's not copy-protected, but it won't copy.
I'd have to say that Toolworks isn't doing a good enough job at playtesting
this program. I think maybe their playtesters are goofing off, if garbage
like this is allowed to be mass-marketed. (Again, I could have a defective
copy, but I've seen other people complain, so I'm doubtful)
Louis Blair
Here is what he says on a lot of subjects:
Chessmaster 2100
`Chessmaster 2000 used software developed by Novag's
Dave Kittinger ... I would like to warn readers about mis-
leading ads by U. S. Chess and Software Toolwoorks
for the "Fidelity Chessmaster 2100". It is claimed to have
"substantially the same program that earned a 2325 rating
in the World Open". ... the Fidelity program is only used in
the soon to be released versions for 68000 based machines
such as Macintosh, Amiga, and Atari ST. Since the 2325
rating was earned with a 20 MHz 68020 processor, it is
reasonable to expect performance close to 2300 if run on the
top Macintosh model which uses a 16 MHz 68030 processor,
assuming the graphics and features don't slow it down
excessively. On a standard Mac or Atari ST it should run
4-5 times slower, which implies a rating around 2100, as the
name suggests. On the other hand, when Software Toolworks
attempted to translate the Fidelity 68000 program to IBM PC
code, they found that it played even weaker than CM2000 IBM
(by Novag's Dave Kittinger and others). So they opted for an
upgrade by Kittinger for the IBM version. ... results imply a
rating for CM2100 of around 2010 on [a superfast 25 MHz 286
based AT clone] or a 20 Mhz 386, or around 1900 on a standard
8 MHz AT, around 1825 on a 10 MHz XT, and around 1750 on
a standard PC or XT. ... As for Apple versions, I would expect
them to play in the 1700s on a standard Apple or perhaps 1900
on an Apple IIgs. Note that the Apple IIgs version is much
weaker than the regular Apple version because all the new
features slow it down excessively, so those who want strong
play on a gs should use the Apple II version.'
Sargon IV
`Sargon IV (IBM version) is a disgrace. ... The program is
supposed to be an IBM adaptation of the Fidelity Mach II program,
but it appears to have been bungled badly. It occasionally makes
illegal moves, sometimes goes into an infinite search and refuses
to move, and plays many silly moves. ... Sargon IV seems weaker
than its predecessor Sargon III, and should never have been
released. The Mac version is apt to be much better since the Mac
uses the same processor as the Fidelity machines so no translation
was needed. I would still expect the CM version to be better as it
was based on the Mach III, a later and better model than the Mach II.'
Zarkov
`Zarkov ... beat CM2100 by 4 1/2 to 3 1/2 at 15" in my test, and on
my 25 MHz AT it went 3-5 against the Fidelity Mach III but only
1 1/2 - 6 1/2 versus Mondial xl. Based on these games I rate it about
70 points above CM2100, but my sample is too small to be very accurate.
The features cannot compare to CM, but in view of its strength and low
price I can recommend it.'
It is odd that Zarkov did better (3 out of 8) against the Mach III (rated by
CRA to be 2265) than it did (1 1/2 out of 8) against the Mondial xl (rated
by CRA at 2154). Of course one would expect a lot of variability in test
results involving such a small number of games. Larry Kaufman has
mentioned that on some of his tests he starts the game from a position
evaluated as even by ECO, instead of starting from the beginning. He calls
this sort of testing reversal testing because for each position two games are
played with each computer getting a chance at the white pieces in order to
ensure fairness. The reason for this sort of testing is to prevent one machine
from getting an advantage over another because of an artificially contrived
opening book. I did not notice any indication of whether or not reversal
testing was used for Zarkov. It is also worth noting that the Kaufman Zarkov
tests were done at the 15 seconds per move speed, while the CRA tests
were done at a much slower speed. Larry Kaufman said, `One point that has
become clear to me is that the choice of time limit makes a significant
difference in the relative ratings. ... For example the Novag Super Expert B
seems to be somewhat stronger than the Fidelity Mach III at Action chess
level, while the reverse is the case at 40/2.' Kaufman mentions the 1989
ACM tournament where slow time controls were used. Here is a rerun
of the results of that tournament as reported by Ken Thompson:
1 Hitech 2400 2513 9+b 6+w 2-w 4+b 5+b 4.0
2 Deep Thought 2551 2508 7+b 8+w 1+b 5+w 3-b 4.0
3 Mephisto X 0 2309 10=w 5-b 6+b 7=w 2+w 3.0
4 Bebe 2150 2169 8-b 9+w 7+b 1-w 10+b 3.0
5 Rebel X 0 2307 6=b 3+w 8+b 2-b 1-w 2.5
6 Cray Blitz 2250 2234 5=w 1-b 3-w 9+b 7+w 2.5
7 Phoenix 0 2053 2-w 10+b 4-w 3=b 6-b 1.5
8 BP 0 2014 4+w 2-b 5-w 10=b 9-w 1.5
9 Novag X 2092 1999 1-w 4-b 10=w 6-w 8+b 1.5
10 Zarkov X 0 1959 3=b 7-w 9=b 8=w 4-w 1.5
A. I. Chess
`A. I. Chess is clearly the strongest pc program on the market at this writing.
... Unfortunately the graphics are horrendous and the price outrageous'.
Again Kaufman seems to have only tested A. I. Chess at fast speeds.
Learning
`The new Elites now have the ability to "learn" from mistakes, but this is only
of slight value as the learning is specific to the position. ... If the
fatal error
occurred many moves before the loss of material the play may be varied too
late to matter.'
Novag Super V. I. P.
`clearly the strongest hand-held unit except for $300+
Mephisto models.'
Fidelity Designer 2100 Display
`has no rival as of yet under $200'
Mephisto Mondial 68000 xl
`still the strongest model under $300 and good value'
Fidelity Designer Mach III
`strongest [model in its] price category'
Novag Super Expert B
`the best wood autosensory buy under $500 considering
strength, features, and board'
Fidelity Elite 2265, Version 2
`the strongest autosensory model under $1000 at this
writing, along with Mephisto Polgar.'
There is a lot more. I think it is easily worth $7.80 to get these
reports four times per year. The address is CCD Inc. c/o I. C. D.
Corp., 21 Walt Whitman Road, Dept. Q, Huntington Station, NY 11746.
Stuart Cracraft
Stuart Cracraft (the well known computer chess expert
and asker of "deep and significant" questions) writes:
>On an IBM PC XT [Sargon 4 and Chessmaster 2100]
>are, at most USCF 1680 and 1800 respectively
and
>AiChess is 300 points higher than Sargon 4, 200 points
>higher than Chessmaster 2100, and about 125 points
>above Zarkov and Psion.
Louis Blair (the well known "chess groupie" who writes
"mass verbiage" and has never written any portion of a
chess program) asks the following not-very-deep questions:
(1) Who did the testing that produced these conclusions?
International Master Larry Kaufman. Estimates are preliminary,
but Larry has erred on the side of lower ratings more often than
on the side of too high.
(2) Did this person once tell us, "The ... score [of a match
recently contested between Fidelity Mach III ... and ZARKOV,
a new chess program available for the IBM PC] was 7 points
for the master machine and 3 points for ZARKOV. This works
out to about 2153-2190 rating points for ZARKOV when running
on this configuration"? What do you think about the correctness
of this calculation?
The same person couldn't have told you that, because the
person who told you the #1 question answer was Larry, and the
one who told you the #2 quote was me. Both are based on similar
testing techniques, and both are involved with computer
chess writing, me actually a little longer than Larry, though
Larry has much more chess talent.
In a standard ELO table, a 70%-30% result is given as
a rating difference of 146-153 points. Extrapolating, mind
you, to a small sample, gives 2265 less 150 points. The first
is Mach 3's CRA rating, and the latter is the ELO table number.
Now, Louis, all you have to do is use your head. No one,
not even your messiah "Larry K." takes the full 150 points as
the difference for computer vs. computer matches. Instead, anywhere
between 50% to 25% is taken, hence the 2153-2190 range.
I'll spell it out for you:
2153 = 2265 - (.75 * 150) = 2265 - 112 = 2153 (wow, it matches?)
2190 = 2265 - (.50 * 150) = 2265 - 75 = 2190 (wow, again?)
The "correctness of the calculation"? Well, please arguee with
the numbers presented above. Ignore the small sample size, or
if you like, make it 14-6 out of 20 games which is what two
Zarkov matches against Mach 3 have produced. So your sample
just increased and gave the same proportions. Harder for you?
(3) Did the test conductor have any financial involvement
in any of the tested programs?
Does your question imply that anyone with any involvement in
a tested program is unlikely to produce non-biased results?
If so, why do you listen to Larry so much? He obviously has
extremely close ties to Novag and Mephisto. Does that invalidate
much of his writings? And what about his strong involvement
with Rex-Chess? Do you write off anyone who is involved
with a chess program? Guess you better write off both Larry
and me.
(4) Do you know who pays for the advertisement in Chess Life
that reads, "Zarkov, expert strength chess, IBM PC and
compatibles. $25-$35, depending on configuration. MSM,
5 Via Amistosa ..."? Do you know how long that advertisement
has been apppearing in Chess Life? Do you know how much,
it has cost to run that advertisement?
How about the other one that reads "Muon Software"? They
also advertise Zarkov and sell the fine program through that
address. And other people are climbing on the Zarkov bandwagon
too. You can't keep good software down, even through misinformation
and mere parroting of data which seems to be your style.
(5) Was the test conductor involved in the writing of the opening
book for any of the tested machines? Had any of the other
programs been used as opponents while the opening book for one
of the programs was being written?
Several of them. And the book is by far the WEAKEST part of
the program! It is really laughable to claim this might bias
the result in FAVOR of the program. The opening book plays
poorly against the massive opening books in Mach III and the
other commercial opening books. It is perhaps 1/10th the
size and prepared by no one near a master-level. Yet they
still do well against them. Even still. I wrote the
primary backbone for the opening book of AI Chess as well.
Both authors (John Stanback of Zarkov and Marty Hirsch of
AI Chess) should well consider their opening books original
with their ammendations and substantive changes.
(6) Is it possible to improve a machine's performance against
specific other programs by making changes in the opening book?
Of course. Just keep tuning it until you find a variation
that wins. Your opening book variation lengthens as you approach
the variation that defeats the opponent machine, or you just find
a refutation to their opening line. Was this done for Zarkov? Not
a chance. See last question and its answer. Is it done by some?
There is documented evidenced it has been done.
(7) How many games were used in the tests? What accuracy
could one expect from rating calculations based on this number
of games?
Now the number is up to 20 games. See answer to an earlier question.
Your argument was stronger when there were only 10 games in the sample.
Guess I just have to keep playing games and building the sample.
You'll need a statistics book to answer your accuracy question.
(8) What time control was used for the tests? Were the tests
between human and machine or between machine and machine?
Are machine-machine tests at one time control an accurate way
to predict what results one could expect between a machine and
a human at a significantly different time control?
Time control was Larry K's standard 60/15.
Stuart Cracraft writes:
>[Zarkov version] 1.30 is now in release
(9) Do you remember this note?
"Subject: NACCC final crosstable
Date: 17 Nov 89 03:42:28 GMT
1 Hitech 2400 2513 9+b 6+w 2-w 4+b 5+b 4.0
2 Deep Thought 2551 2508 7+b 8+w 1+b 5+w 3-b 4.0
3 Mephisto X 0 2309 10=w 5-b 6+b 7=w 2+w 3.0
4 Bebe 2150 2169 8-b 9+w 7+b 1-w 10+b 3.0
5 Rebel X 0 2307 6=b 3+w 8+b 2-b 1-w 2.5
6 Cray Blitz 2250 2234 5=w 1-b 3-w 9+b 7+w 2.5
7 Phoenix 0 2053 2-w 10+b 4-w 3=b 6-b 1.5
8 BP 0 2014 4+w 2-b 5-w 10=b 9-w 1.5
9 Novag X 2092 1999 1-w 4-b 10=w 6-w 8+b 1.5
10 Zarkov X 0 1959 3=b 7-w 9=b 8=w 4-w 1.5"
[Blair: Which version of Zarkov was this? Is 1959 "expert strength"?]
And a sample size of 5 is more reliable than the sample size of
20 quoted above? Is this a real scientific argument you're constructing
or merely acting the gadfly?
Stuart Cracraft writes:
>DT's surprising 29. Qg5 against Cooke at the Software
>Toolworks is apparently not too difficult to find, at least
>for a machine. The PC program "ZARKOV" found and
>played 29. Qg5 in only 5 ply of search (not counting
>extensions for captures and checks). The winning Qg5
>was found at the second to last variation to be searched
>at the 5th ply, in 4 minutes 30 seconds, on a "nothing great"
>16mhz 80386.
(10) Why stop with Zarkov? Why not see how low we can
go? (I tried it on the approximately five-year-old Novag
Super Constellation at level five (approximately two minutes
per move). It played 29 Qg5 after almost exactly two minutes.)
Does anyone have Sargon III or one of those cheap Radio
Shack machines?
Super Constellation is optimized for tactics, not positional play.
Of course it finds that move fast. Any Kittinger program would.
Tell me if Sargon 3 or "cheap Radio Shack" (you libel SciTek)
can solve it.
(11) How much agreement is there between the ratings
posted by Stuart Cracraft and the opinions of Larry
Kaufman, primary author of Computer Chess Reports?
The latest issue will soon be available.
Larry certainly has complimented AI Chess and Zarkov. He actually
SAW AI Chess achieve a 2360 performance rating, running on an
IBM PC 25mhz 386, at software toolworks -- while the massive, 10 year
long constructed Mephisto program BARELY achieved 2390 something.
And AI Chess, is less than 2 years old PART TIME, by one person
without IM or GM input and only two tournaments.
Louis, you should be talking about AI Chess -- not Zarkov.
--Stuart
Dennis J. Kosterman
|
|In article <29...@astroatc.UUCP> stu...@astroatc.UUCP (Dennis J. Kosterman) writes:
|>
|> 1. Some of us (like me) are mediocre chess players. My best USCF
|>rating (when I had one) was 1328, and it'll be years before I reach 1800,
|
|Ummm, I'm not sure I agree with this attitude....
| ... My attitude had always
|been that I'd rather lose a good game that I could learn from and
|show people than bungle my way to a win.
I basically feel the same way, except that you need to win once in
a while or you start to lose your confidence and enthusiasm for the game.
But over the long run, I think you're right. I'd prefer to play somebody
higher-rated than myself -- it seems to make me play better, and I can
learn more from the game, win or lose.
But how much higher-rated? I don't think I'd get much out of losing
to a grandmaster, unless he or she was kind enough to go over the game
with me, and even then I'm not sure. The grandmaster is playing at so
many levels above me it's almost a different game. His/her reasons for
doing things might not even make sense to me.
My conclusion is that the optimum rating for an opponent, in terms of
learning from the game, is about 200 points above your own rating. Much
higher than that, and the game becomes too one-sided for you to learn much.
It also keeps the game competitive -- not so that you always (or even usually)
win, but so that you always at least feel that you *could* win. If it's hope-
less right from the start (e.g., me vs. a grandmaster, or even an expert),
I lose interest.
Dennis J. Kosterman
uwvax!astroatc!stubbs
Tom Friedel
>power, and openings. I've been told that most of
>this was written by Tom Friedel. The rest of the
What?????? NO !
tom
Tom Friedel t...@jdyx.uucp (404) 320-7624 BBS: (404) 321-5020
Stuart Cracraft
>If a person can afford it, I often suggest that a person
>buy a cheap chess computer and use it for awhile before
>putting money into an expensive machine. After some
>experience with a chess computer, a person can have
>a better idea about what features are important to him.
If cost is a factor, then a person shouldn't lay out ANY money.
If they're on usenet, they very likely have access to a
Unix machine and can beg, borrow, or steal GNU Chess 1.55.
>A cheap Saitek machine might be good for this purpose.
>(No, Stuart, that is not "libel". I am sure that anyone in
>Saitek would agree that they sell cheap machines and
>expensive machines. I am simply suggesting one of the
>cheap ones.)
>
But that gets my consumer advocate hair-on-the-back up again.
Saitek or otherwise. Fidelity is really thought of as the
premier supplier of inexpensive machines. However, if you
can do it for FREE on your existing Unix machine, don't
buy a chess machine. They are antiquated, ridiculously
expensive, especially at upgrade time, and are rapidly dying
as newer, free software like GNU Chess and IBM PC software
like AI Chess and Zarkov become more widely known. Even the
overpriced AI only costs $19.95 to upgrade. Why bother with
Mephisto's and others that charge hundreds or even thousands
for upgrades. Or why bother buying a $300 machine when you know
next year's will cost the same to upgrade to?
>I have not seen much information on the playing strength
>of the under-$100 chess computers. If one is determined
>to stay under $100 (thus ruling out things like the Novag
>Super V. I. P.), the Saitek machines may be about as good
>as anything else. Does anyone have more information on
>machines in this price range?
Once you eliminate the cost, you eliminate the need to have
a consumer advocate. Then it just becomes one place's
free chess program vs. another. So if you have an IBM PC, get
Psion since it appears to be in the public domain due to a
misrelease to the public 5 years ago. If you have access
to a faster machine running Unix, get GNU Chess 1.55. If
you want the strongest program without regard to cost, get
AI Chess.
Stamp out consumer advocacy!
--Stuart
Elliott Winslow
The USCF over-zealously fell for the conclusion that the Saitek PR
people wanted the consumer to fall for, namely think that Kasparov
actually wrote the accompanying book with the Cavalier. The only
input he had was that he wanted to include some game of his. Oh, well.
Frederic Friedel of West Germany and ChessBase was contracted to write
the manual for this machine (and for much of Saitek's documentation), and
I was little by little asked to come up with games (with Julio Kaplan),
suggest quiz positions, comment a little on them, then some more, until
it had turned into a whole tutorial. No royalties, and it was pressed
into my schedule from the side, so to speak, but it was fun. Frederic
(I'm guessing here, but with a high likelihood of truth) had the last
pass at it. So my greatest pun, after 10. Nxb5 in Morphy vs. Count
Isouard & the Duke of Brunswick, was originally "The only regal move!",
playing off "only legal," and, as it happened, playing off standard
mis-pronounciation (believe me, this was an accident!) -- and in the
book it appears as "A truly regal move." I can't decide if this is
better or worse.
In any case, I think Frederic did a splendid job, and wonder if he is
available to write the manual for ChessBase!?
By the way, I hope to be a bit more candid on here than you may have
expected from the USCF, with the appropriate employer (current and ex-)
disclaimers, etc. etc. Just bear with me, I post infrequently. E-Mail
readily accepted.
{apple,pacbell,hplabs,ucbvax}!well!ekw
Elliott
Louis Blair
>I've been told that most of
>this was written by Tom Friedel.
Tom Friedel writes:
>What?????? NO !
Oops. Sorry, Tom. I should have
said, "FREDERIC Friedel".
For the benefit of Stuart:
Yes, world, I Louis Blair was all
wrong about Tom Friedel.
Louis Blair
prices. I'M not selling anything. Anyway,
I was interested in this remark by Stuart
Cracraft:
>Fidelity is really thought of as the
>premier supplier of inexpensive machines.
We get a LOT of questions about chess com-
puters for under $100. Fidelity, Novag, Saitek,
and probably others all sell machines in this
price range. My guess is that Saitek machines
are the weakest of the three in terms of playing
strength, but that is just a guess. It would be
nice to have more solid information. Does anyone
have it? If not perhaps it would be a good idea
to have some public games on the net between
these machines. Watching them bumble around
might be fun and instructive to those who would
consider buying a machine in this price range. For
this purpose, I can get access to a $75 portable
Saitek machine. If anyone has a portable machine
in the same price range from a rival manufacturer,
then we can see how they do against each other.
My guess is that Saitek will lose, but who knows.
I would like to see a game between a Fidelity
machine and a comparable Novag machine.
Stuart Cracraft
>I don't know why Stuart talks about MY
>prices. I'M not selling anything.
Except occasional disinformation, lots of personal invective,
and semi-accurate advice maybe?
(Letting bygones be bygones...)
>I was interested in this remark by Stuart
>Cracraft:
>>Fidelity is really thought of as the
>>premier supplier of inexpensive machines.
>
Gawd how I like to see myself quoted!
>We get a LOT of questions about chess com-
>puters for under $100. Fidelity, Novag, Saitek,
The "Royal We" it sounds like...
>and probably others all sell machines in this
>price range. My guess is that Saitek machines
>are the weakest of the three in terms of playing
>strength, but that is just a guess. It would be
Easy guess.
>nice to have more solid information. Does anyone
>have it? If not perhaps it would be a good idea
>to have some public games on the net between
>these machines. Watching them bumble around
Why bother. Get Larry's digest. So much time
spent re-inventing the wheel...
>My guess is that Saitek will lose, but who knows.
>I would like to see a game between a Fidelity
>machine and a comparable Novag machine.
Nowadays at the low level, it's probably
Saitek-Fidelity-Novag with the last coming first.
As for Mephisto, might as well write them off
in this category...
Novag has become impressive of late. Sadly the
lack of a CRA rating holds back their sales,
even though the basic program is probably stronger
than Fidelity's.
Some news: all 20 of Mephisto's $10,000 machines
have sold out. It is amazing what people will
spend on -- when they could just slap AI Chess
1.34 on their home 25mhz 286/386 and end up having
the same 2400 rating....
Stuart
Louis Blair
>I'M not selling anything.
Stuart Cracraft writes:
>Except occasional disinformation,
>lots of personal invective,
>and semi-accurate advice maybe?
Of course, I deny this. Any disinfor-
mation, personal invective, and semi-
accurate advice that I post comes
completely free of charge.
I wrote:
>perhaps it would be a good idea
>to have some public games on the net between
>these machines.
Stuart Cracraft writes:
>Why bother. Get Larry's digest. So much time
>spent re-inventing the wheel...
I had in mind machines like Novag Solo Chess,
Saitek Pocket Chess, Fidelity Mini Chess, Fidelity
Micro Chess, Fidelity Chess Card, Novag Secondo,
Mephisto Mini, Saitek Cavalier, Fidelity Gambit, Saitek
Electronic Chess Partner, Fidelity Genesis, Saitek
Mark 12, Fidelity Excellence, Fidelity Chess Coach 1500,
and Novag Mentor 16. That is, I was wondering about
machines that are at the bottom of the price spectrum.
Only one of these is in Kaufman's rating list and Kaufman
indicates that he used a "small sample" for this rating.
Unfortunately, it appears that a lot of people are determined
to buy machines like this. I thought it would be nice to
see how they compare with each other and get a general
idea of just how low the quality of play is for machines
like this. Of course, a future issue of Kaufman's report
may have more on these machines, but it might be fun
to do this while we wait.
Stuart Cracraft writes:
>people ... could just slap AI Chess
>1.34 on their home 25mhz 286/386 and end up having
>the same 2400 rating....
This 2400 rating estimate is based on how many games?
Incidently, has the amount of your financial involvement
with the selling of AI Chess changed since the last time
you talked about your connection with this product?
GaRY NEwelL
> This 2400 rating estimate is based on how many games?
> Incidently, has the amount of your financial involvement
> with the selling of AI Chess changed since the last time
> you talked about your connection with this product?
Come on -like ol Stu is going to give us the truth here right?? In the two
years that I've read this group, Cracraft has acted like some kind of
anal-retentive, self-righteous ass and has consistently posted articles
that are essentially advertisements for whatever venture he is currently
involved in - without mentioning his financial interest in the product.
But I still think that the time he had the audacity to contact the academic
advisor of a poster to this group and attempt to get this person into hot
water because he didn't like something that was said was the ultimate indicator
of the type of person that Stu cracraft is.
I for one hope that either you will continue to take him up on his
self-serving claims and force him to back them up - or that he will
take enough rope and hang himself with it as he has done before.....
gary newell
Roy Mongiovi
> I for one hope that either you will continue to take him up on his
> self-serving claims and force him to back them up - or that he will
> take enough rope and hang himself with it as he has done before.....
I, for one, wish that both of them would take their feud to mail, where it
belongs, and only post when they have something to say about chess. I decided
Cracraft was a pompous ass all by myself a long time ago. I don't need Blair
to keep pounding it into me again and again.
Roy J. Mongiovi Systems Support Specialist Office of Computing Services
Georgia Institute of Technology Atlanta, Georgia 30332-0275 (404) 894-4660
uucp: ...!{allegra,amd,hplabs,ut-ngp}!gatech!prism!roy
ARPA: r...@prism.gatech.edu
Sohan C. Ramakrishna-Pillai
Re: .. Roy Mong...@prism.gatec (820)
> > In article <16...@megaron.cs.arizona.edu>, g...@cs.arizona.edu (GaRY
> NEwelL) writes:
> > I for one hope that either you will continue to take him up on his
> > self-serving claims and force him to back them up - or that he will
> belongs, and only post when they have something to say about chess. I decided
> Cracraft was a pompous ass all by myself a long time ago. I don't need Blair
> to keep pounding it into me again and again.
> Roy J. Mongiovi Systems Support Specialist Office of Computing Services
In a way, I find the incessant exchange boring too, but one has to
remember that while many of us old-timers on this newsgroup remember the
time when Mr. Stuart Cracraft's machinations were exposed by Mr. Louis
Blair, new readers may not do so and might be taken in by the
authoritative style Mr. Cracraft has arrogated to himself to use when it
comes to peddling his wares or asserting his self-assigned rights as a
founder/co-founder of this newsgroup.
So, Mr. Blair, until Mr. Cracraft shows indications of reform, here's
another vote for you (rather, a request to you) to keep pounding on his
misleading (and sometimes, downright dishonest!) articles. You are doing
potentially gullible readers a great service by trying to keep Mr.
Cracraft's salesmanship honest.
For bored readers, please note that the exchanges are REACTIVE on Mr.
Blair's part and are caused by Mr. Cracraft's posting of misleading
articles and Mr. Blair's demand for greater honesty/accuracy on Mr.
Cracraft's part. The best way to stop this is by censuring Mr. Cracraft,
rather than wishing such exchanges away or being intimidated by Mr.
Cracraft's aggressiveness into acquiesing into silence.
P.S. My relationship with Mr.Blair is mostly professional - he is the
local chess club's (ex-?)librarian - plus the minor acquaintance this
relationship results in. I have no acquaintance with Mr. Cracraft beyond
reading his postings on this newsgroup.
___
Sohan C. Ramakrishna-Pillai
Office: UCC 181 Phone: x6406
Home: Phone: (412)661-9375
Inverse Majority Belief Law: The likelihood of the correctness
of a belief is inversely proportional to the number of people holding
it.
Tim Thomas
In article <48...@hydra.gatech.EDU> r...@prism.gatech.edu (Roy Mongiovi) writes:
>> In article <16...@megaron.cs.arizona.edu>, g...@cs.arizona.edu (GaRY NEwelL) writes:
>> I for one hope that either you will continue to take him up on his
>> self-serving claims and force him to back them up - or that he will
>> take enough rope and hang himself with it as he has done before.....
>
>I, for one, wish that both of them would take their feud to mail, where it
>belongs, and only post when they have something to say about chess. I decided
I couldn't agree more.
--------------------------------------------------------------------------
Tim Thomas t...@shire.cs.psu.edu *!psuvax1!shire!tim
Stuart Cracraft
>Come on -like ol Stu is going to give us the truth here right?? In the two
And like good ol' Gary is going to put anything into the
newsgroup that's chess-related? Honestly, you say the same
things you most despise, or say you do, in hearing another
poster say!
At least people can challenge tests. Though Blair's challenge
of a 2400 rating from a short test seemed very short, terribly
unscientific. Then, when I reply a little strongly to the
collegian, the rest kick in.
Kids, it's not football -- it's chess.
Stuart
P.S. Soon, you'll have to fend all for yourselves.
Get your licks in now, when it's fun!!!
Mark E. Glickman
> Why bother. Get Larry's digest. So much time
> spent re-inventing the wheel...
That is, if you trust Larry Kaufman wholeheartedly. Maybe you don't
know, but Larry Kaufman used to be the chairman of the rating committee.
One day, an advertisement for one of the new chess computers appeared in
Chess Life which quoted
him as saying that this particular computer had such and such a rating.
It was pretty clear that Kaufman was paid to contribute to the
ad.
He was promptly kicked off the rating committee. He did not voluntarily
withdraw - he was disspelled. I wouldn't trust him with my kids
(well, okay, I don't have any kids).
It's the same way it's hard to trust Stuart Cracraft each time
he claims that Zarkov is the best PC-based chess program without
mentioning his monetary involvement.
- Mark
Nick Lai
>One day, an advertisement for one of the new chess computers appeared in
>Chess Life which quoted him as saying that this particular computer had
>such and such a rating. It was pretty clear that Kaufman was paid to
>contribute to the ad.
Just because an ad quotes an "expert" does not mean that said "expert" was
paid for the accolade. How many movie ads scream about their "enthusiastic
thumbs up from Siskell && Ebert"? When was the last time you looked at
the back of a best-selling paperback? You'll find a bunch of quotations
(sometimes out-of-context) of reviews that appeared in various newspapers,
magazines, and TV shows (and sometimes from friends of the author!).
If you have some credible proof that Kaufman was ejected from the
ratings committee for some wrongdoing, please post the text of said proof
and give a complete citation. Otherwise, we can only assume that your
posting in slander.
Nick
Louis Blair
>Larry Kaufman used to be the chairman of
>for one of the new chess computers appeared in
>Chess Life which quoted him as saying that this
>particular computer had such and such a rating.
>It was pretty clear that Kaufman was paid to
>the rating committee. He did not voluntarily withdraw -
>he was disspelled. I wouldn't trust him with my kids
>(well, okay, I don't have any kids).
>It's the same way it's hard to trust Stuart Cracraft
>each time he claims that Zarkov is the best PC-based
>chess program without mentioning his monetary
>involvement.
Nick Lai writes:
>If you have some credible proof that Kaufman was
>ejected from the ratings committee for some wrongdoing,
>please post the text of said proof and give a complete
>citation. Otherwise, we can only assume that your
>posting in slander.
With a constant blast of commercials in our ears (well, eyes
actually), I think we all get a little bit short-tempered. Here,
I think there is a little bit of truth on both sides. Here is what
Kaufman himself had to say on the subject in his 1986 Compu-
ter Chess Reports. (He had just finished describing some very
questionable aspects of the CRA Par Excellence 2100 rating
test.)
"At the time of these rating tests I was the chairman of
the (human) ratings committee of the USCF. Although computer
ratings are handled by a separate committee so my committee
had no jurisdiction, once I learned the above facts, I made
my objections known very clearly. Because of my prior
endorsement of the Munich `S' (made voluntarily as a
customer of ICD and before I became affiliated with
`Computer Chess Reports'), I was perceived by USCF as biased
against Fidelity. Rather than address the issues I raised,
it was made clear to me that my comments were unwelcome.
Fidelity volunteered to make available to me a pre-
production model of the `Par' for testing, ... I made all results
available to anyone who asked, as there was no discussion of
confidentiality and no apparent reason for it, as the model I tested
was identical to the commercial version, except for one opening
move which Fidelity changed on my advice ... When
ICD attempted to use my early test results in an ad, not
only was the ad refused, but I was relieved of my post on
the ratings committee on grounds that my work on this
`Computer Chess Reports' was a conflict of interest with my
ratings post, even though the ratings committee did not deal
with computer ratings. The irony is that the USCF official
who made this determination of conflict of interest is in
charge of product sales!"
A year later, CCR reported, "Larry has now been reinstated on
the human ratings committee".
In short, it does appear that Kaufman was at one time
"ejected from the ratings committee" because of CLAIMS of
"wrongdoing", but I think there is good reason to question
exactly who was doing the wrongs.
As I have said before, people should not completely trust ANY
one source for chess computer rating information. People should
try to get information from as many different sources as possible.
It must be remembered that Kaufman does have some financial
involvement in the business himself: He writes his reports
(although I doubt that they are very profitable), makes occasional
contributions to the opening books of various machines, and is
now involved in trying to sell his own software, Rexchess. Within
limits, I am inclined to trust him. On those occasions when I have
tried to confirm what he said, everything checked out. In his
reports he DOES mention his own financial involvement, innac-
curacies in test results due to small numbers of games, problems
arising from different time controls, etc.
ed.knowles
> In article <58...@rice-chex.ai.mit.edu>, crac...@wheaties.ai.mit.edu (Stuart Cracraft) writes:
> > Why bother. Get Larry's digest. So much time
> he claims that Zarkov is the best PC-based chess program without
> mentioning his monetary involvement.
>
> - Mark
Some time ago (a year, maybe) I challenged Stuart to send me a copy
of AIChess to test against other programs on various machines. He
declined, but John Stanback promptly sent me a copy of Zarkov. He
has been patiently awaiting my test results -- I haven't completed
everything, and time constraint and machine availability problems
are holding me up. John has continued sending upgrades nevertheless.
I hope to reward his patience eventually with an evaluation noone
can find technical fault with. However, I feel moved to make some
comment about Zarkov, since I feel that Stuart may have unwittingly
slighted the program and indirectly "hurt" John. My only financial
interest in Zarkov is my free copy and upgrades that John sends me.
My rating over the last few years has ranged from a low of 1896
(June 1988) to a high of 2095 (February 1987) -- my current rating
is 1932. MY PERSONAL IMPRESSION of Zarkov is that at fast time
controls (60/5, 60/15, 60/30, 60/60), Zarkov is considerably weaker
than PSION chess and slightly weaker than CHESSMASTER 2000 (CHESS-
MASTER 2100 doesn't seem any stronger than CHESSMASTER 2000). At
slower time controls (60/120, 40/120, 40/240 etc...), there seems
to occur a quantum leap in strength, and Zarkov so far seems superior
to either program. For some reason, PSION doesn't seem to play
much better at level 6 than it does at level 3. Zarkov beats me
about 60% of the time at 60/120 while I beat it quite easily at
time limits faster than 60/60. So at the slower time limits , I
feel that the program plays at least my strength (~1900). For
fast games I prefer PSION, for slow games I prefer ZARKOV, and
for the same reason -- they're a better challenge. I never use
CHESSMASTER 2000/2100 except in testing ZARKOV.
Realize please, that this is a subjective evaluation. When
I compile and analyze the dozens of games I've played against
ZARKOV (using myself, PSION, and CM2000 as opponents), and sent
them off to John Stanback, I'll post a more objective evaluation.
But I must adamantly agree with Stuart that ZARKOV is a great
program for the price. I like it much better than CM2100.
The opening book for ZARKOV is pretty bad (sorry, Stuart), but
a great feature of the program is that you can add to the opening
book and create your own.
These are my honest opinions, biased only by the fact that I
received ZARKOV free. I post this without the permission of John
Stanback (the creator of Zarkov) or Stuart Cracraft.
-Ed Knowles
Patrick L Faith
(85$) one game against Sargon 4 on my Mac SE with the Kasparov winning due to
a poor move by Sargon on about the 11th move(Sargon was to agressive). I
use the Cavalier as a portable chess set which happens to have a computer in
it. I think anyone with a rating above 1600 can consistently win the
cavalier at the 10 second per move mode.
P L FAITH
Louis Blair
"free book by Kasparov" with every
Saitek Kasparov Cavalier that they
sell. The latest catalog says, "Free
190 page book `Kasparov Chess' with
each order!" Elliott, was this your doing?
Elliott C Winslow
Yes. The people in advertising are actually interested in telling the
truth, to their credit.
I just saw a new Saitek catalog -- that "coach" feature is now in
two or three machines. Oh, if I had just demanded royalties! I'd be ...
well, I'd be less poor.
--Elliott
Louis Blair
>In article lb...@andrew.cmu.edu (Louis Blair) writes:
>>USCF no longer claims to provide a
>>"free book by Kasparov" with every
>>Saitek Kasparov Cavalier that they
>>sell. The latest catalog says, "Free
>>190 page book `Kasparov Chess' with
>>each order!" Elliott, was this your doing?
>
>Yes. The people in advertising are actually interested in telling the
>truth, to their credit.
This certainly IS an improvement, but I
can not resist noting that "190 pages"
is a little misleading since three fourths
of the book is in various foreign languages.
Or am I being too picky?