M-Chess 7.0 - Nice!
john quill taylor
I've been playing with M-Chess 7.0 for about a week now,
and it's a very nice program for under US$50!
During fast (2 sec/move) play, however, I have experienced
one irritation: once in a while, if I click the piece to move,
and then click the destination square, the move is not made.
Upon redo, it always works. Is this a bug, or does it happen
on some machines and not others? Is there a patch or a fix?
Another minor quirk is when it finds, for example, "Mate in 13"
under "Infinite" search, a very short time later it often says,
"Mate in 12." I suspect that at this point, it's LOOKING for a
shorter mate, but it displays the lower number when it shouldn't.
This is a very minor issue (who ever thought MICRO computers
would even be dealing with announcing Mate in 13? And this
example was NOT using the tablebases!)
Other than these minor complaints, M-Chess 7.0 is definitely
keeping in the tradition of robust, well-behaved piece of chess
software. In very rare instances, it takes a few seconds
longer than M-Chess 5.0 to find some tactical moves, but I
think this is the overhead from the increased hash size.
More often, I've noticed that it finds moves in minutes that
earlier versions could never find, even overnight.
The tablebases are great (though they tear your heart out
when you see it announce "Mate in 12" in an endgame that
you thought you could draw!) I notice the "x" mode has gone
away, so it's a bit easier to run M-Chess 7.0 in DOS, Win 95,
and Win NT (exit to DOS from Win 95 has EXCEPTIONAL speed,
so a DOS boot disk is really no longer necessary for optimum
performance). All in all, an excellent job, Mr. Hirsch!
- jqt -
Komputer Korner
M-Chess has found the mate in 13 by some extension sequence but the
later finds the shorter mate. All programs will exhibit this
behaviour.
--
--
Komputer Korner
The inkompetent komputer
To send email take the 1 out of my address. My email address is
kor...@netcom.ca but take the 1 out before sending the email.
john quill taylor wrote in message
<6gte45$h2...@hpbs1500.boi.hp.com>...
john quill taylor
"Komputer Korner" <kor...@netcom.ca> wrote:
>M-Chess has found the mate in 13 by some extension sequence
>but [then] later finds the shorter mate. All programs will
>exhibit this behaviour.
Yes, but this seems like a (minor) bug. Here is my test:
file: g6-wins
date: 14 April 1998
item: M-Chess Pro 7.0 Mate-in-16 Analysis
machine: HP OmniBook 3000
cpu: Pentium II 266 MHz MMX
memory: 80 Megabytes SDRAM
program: M-Chess Pro 7.0
mode: 3 (default settings used)
hash: 60 Megabytes (maximum allowed)
boot: DOS Command Prompt Only [F8]
other: using HIMEM.SYS (as recommended)
position: White to move... Mate-in-16 (+M16)
White: Kc7, Pa5, c2, d5, e5, g5
Black: Ka8, Pa6, b3, c4, e6, f7, h3
line: 1.g6!! fxg6 2.d6 b2 3.d7 b1=Q 4.d8=Q+ Ka8 5.Qd4+ Ka8 6.Qe4+ Ka7
7.Qe3+ Ka8 8.Qf3+ Ka7 9.Qf2+ Ka8 10.Qf8+ Ka7 11.Qc5+ Ka8 12.Qc6+ Ka7
13.Qb6+ Qxb6+ 14.axb6+ Ka8 15.b7+ Ka7 16.b8=Q#
move: 1.g6!!
found: 04:31
depth time nodes value
13-2/11 04:31 09.8M +0.01
14-1/11 14:44 32.4M +0.61
14-1/11 21:40 47.2M +1.51
14-1/11 21:55 47.8M +M16 (mate in 16)
16-2/11 22:35 49.5M +M15? (bug?)
Miscellaneous Items Noted:
Mate in N-1 Announced
---------------------
M-Chess 7.0 displays "mate in 16" (correct) and then later displays
"mate in 15" (incorrect). Usually, one or two plies after the proper
"mate in X" gets detected, the program erroneously decrements the
announced "mate in X" by one move. This seems to happen at only
"mate in 12" and higher.
Refresh Old Evaluation Value Appears
------------------------------------
Refresh (+) resets the evaluation (Value) from +0.61 back to the
previous value of +0.01 (incorrect).
Mate in 4 Auto Move
-------------------
At Mate-in-4 or lower, program makes the move, even when in "infinite"
search mode. This is probably done intentionally.
Mouse Miscue
------------
When playing blitz (5 seconds per move), occasionally I click to move
a piece, and when I click to show where to move the piece, it ignores
my selection and does nothing, almost as if I had made an invalid move
(which I did not). Upon re-doing the move again, it always works fine.
This may be due to my touchpad mouse; they are notorious for this.
- jqt -
john quill taylor
john-qui...@hp.com (john quill taylor) wrote:
>Yes, but this seems like a (minor) bug. Here is my test:
The mouse trouble (last item below) I had with M-Chess seems
to be confined to one particular machine; I should have done
more testing before posting this particular problem.
It is a problem with my hardware, not an M-Chess problem.
The other items can be duplicated across different PCs...
>Miscellaneous Items Noted:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
That's what this is I think, a hardware problem.
- jqt -