Rnd2, CM9k vs Fritz 9k - 1 attachment

Main

Date: 12 Mar 2006 13:29:54
From: Wilma
Subject: Rnd2, CM9k vs Fritz 9k - 1 attachment

Attached is Round 2 of my Chessmaster 9000 versus Fritz9 match. In this
round, the programs were run separately. CM9k, playing white, won handily
with a Ruy Lopez. For those who don't know, CM9k won Round 1 with black. In
both games, CM9k evaluations correlated more consistently with the results
than Fritz did.

Two rounds is enough for me. It's just too much trouble, especially when
the results have been so clear.

Wilma

begin 644 Chessmaster 9000 vs Fritz9 2006.3.9.PGN
M6T5V96YT(")#33EK($9R:71Z.2!-871C:")=#0I;4VET92`B4F%M/3(U-4UB
M(#$N-D=H(%)U;B!S97!A<F%T96QY(ET-"EM$871E("(R,#`V+C,N.2)=#0I;
M4F]U;F0@(C(B70T*6U=H:71E(")#:&5S<VUA<W1E<B`Y,#`P(ET-"EM";&%C
M:R`B1G)I='HY(ET-"EM4:6UE0V]N=')O;"`B-#`O,3(P(ET-"EM297-U;'0@
M(C$M,")=#0I;14-/(")#.#<B70T*#0HQ+F4T(&4U(#(N3F8S($YC-B`S+D)B
M-2!A-B`T+D)A-"!.9C8@-2Y/+4\@9#8@-BY293$@0F4W(#<N8S,@3RU/(#@N
M9#0@#0I"9#<@.2Y.8F0R(&5X9#0@,3`N8WAD-"!.8C0@,3$N0F(S($YD,R`Q
M,BY293,@3GAF,B`Q,RY+ >&8R($YG-"L@,30N2V4R(`T*0F(U*R`Q-2Y.8S0@
M9C4@,38N830@3GAE,R`Q-RY+ >&4S($MH."`Q."YA>&(U(&9X930@,3DN2WAE
M-"!D-2L@,C`N2V4S(`T*9'AC-"`R,2Y" >&,T(&%X8C4@,C(N4GAA."!1>&$X
M(#(S+D)X8C4@46$U(#(T+D)D,R!1:#4@,C4N2V0R(&,U(#(V+F1X8S4@#0I1
M >&,U(#(W+DME,B!"9#8@,C@N46,R(%%B-"`R.2Y+9#$@0F8T(#,P+D)X:#<@
M0GAC,2`S,2Y+ >&,Q(%)X9C,@,S(N9WAF,R`-"E%E,2L@,S,N460Q(%%E,RL@
M,S0N2V,R(%%C-2L@,S4N2V(Q($MX:#<@,S8N46,R*R!1 >&,R*R`S-RY+>&,R
5($MG-B`-"C,X+DMD,R`Q+3`-"@T*
`
end

Date: 16 Mar 2006 17:53:40
From: George
Subject: Re: Rnd2, CM9k vs Fritz 9k - 1 attachment

>If you ran both engines against each other on the same machine, then you
>are not performing a valid test, especially if your machine does not
>have multiple CPUs. The results you are seeing may have less to do with
>the chess engines and more to do with the default configuration of each
>of them allowing one to grap more CPU time than the other.

Not exactly true. It would be completely true if "Pondering/Permanent
Brain" was enabled during these games. But if it is not enabled, then
everything is OK and the match is fair and the results valid.....

Date: 12 Mar 2006 17:09:00
From: George
Subject: Re: Rnd2, CM9k vs Fritz 9k - 1 attachment

>Two rounds is enough for me. It's just too much trouble, especially when
the results have been so clear.

Obviously you don't have any idea about Chess-engines testing.....

Please follow the following rules before you post another one:
1)Post the time controls you used for the match.
2)Post the hash table memory you gave to each program for the match.
3)Post the hardware the engines run.
4)Post the opening books each engine had.
5)Post the games(I see you gave an attachement but it appears as a
complete mess to me).
6)Start making comments about the performace of each engine after 20
matches and start making serious comments after 50 games. But even
after 50 games you should not look very sure about the result. After
100 games you should be able to give with confidence a clear opinion
about engine's strength. Even then, after 100 games, if the result is
close enough, then you should play even more games....

Date: 13 Mar 2006 03:26:27
From: Wilma
Subject: Re: Rnd2, CM9k vs Fritz 9k - 1 attachment

I wasn't doing engines testing. I was seeing who would win a 2-game match.
It was for fun. CM9k won both game, and I reported it.

Don't know why the attachment didn't come through for you. Here it is in
text form:

[Event "CM9k Fritz9 Match"]
[Site "Ram=255Mb 1.6Gh Run separately"]
[Date "2006.3.9"]
[Round "2"]
[White "Chessmaster 9000"]
[Black "Fritz9"]
[TimeControl "40/120"]
[Result "1-0"]
[ECO "C87"]

1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6 5.O-O d6 6.Re1 Be7 7.c3 O-O 8.d4
Bd7 9.Nbd2 exd4 10.cxd4 Nb4 11.Bb3 Nd3 12.Re3 Nxf2 13.Kxf2 Ng4+ 14.Ke2
Bb5+ 15.Nc4 f5 16.a4 Nxe3 17.Kxe3 Kh8 18.axb5 fxe4 19.Kxe4 d5+ 20.Ke3
dxc4 21.Bxc4 axb5 22.Rxa8 Qxa8 23.Bxb5 Qa5 24.Bd3 Qh5 25.Kd2 c5 26.dxc5
Qxc5 27.Ke2 Bd6 28.Qc2 Qb4 29.Kd1 Bf4 30.Bxh7 Bxc1 31.Kxc1 Rxf3 32.gxf3
Qe1+ 33.Qd1 Qe3+ 34.Kc2 Qc5+ 35.Kb1 Kxh7 36.Qc2+ Qxc2+ 37.Kxc2 Kg6
38.Kd3 1-0

Both machines were set to play their best games. Hash tables: Fritz9
default is 37MB, and CM9k default is 4 MB. The machine ws Dell Dimension
4300.

Wilma

Date: 13 Mar 2006 02:25:13
From: Wilma
Subject: Re: Rnd2, CM9k vs Fritz 9k - 1 attachment

There is one item of relevant information you mention that I did not
include in my post or in the attachment heading: the hash tables. I used
the default settings that came with the programs. That's 37MB for Fritz9
and 4 MB for CM9k.

Wilma

\
"George" <gtsavdar@auth.gr > wrote in news:1142212140.215579.228590
@e56g2000cwe.googlegroups.com:

>>Two rounds is enough for me. It's just too much trouble, especially when
> the results have been so clear.
>
> Obviously you don't have any idea about Chess-engines testing.....
>
> Please follow the following rules before you post another one:
> 1)Post the time controls you used for the match.
> 2)Post the hash table memory you gave to each program for the match.
> 3)Post the hardware the engines run.
> 4)Post the opening books each engine had.
> 5)Post the games(I see you gave an attachement but it appears as a
> complete mess to me).
> 6)Start making comments about the performace of each engine after 20
> matches and start making serious comments after 50 games. But even
> after 50 games you should not look very sure about the result. After
> 100 games you should be able to give with confidence a clear opinion
> about engine's strength. Even then, after 100 games, if the result is
> close enough, then you should play even more games....
>
>

Date: 14 Mar 2006 16:39:54
From: Thomas T. Veldhouse
Subject: Re: Rnd2, CM9k vs Fritz 9k - 1 attachment

Wilma <wilma_defemina@hotmail.com > wrote:
> There is one item of relevant information you mention that I did not
> include in my post or in the attachment heading: the hash tables. I used
> the default settings that came with the programs. That's 37MB for Fritz9
> and 4 MB for CM9k.
>

If you ran both engines against each other on the same machine, then you
are not performing a valid test, especially if your machine does not
have multiple CPUs. The results you are seeing may have less to do with
the chess engines and more to do with the default configuration of each
of them allowing one to grap more CPU time than the other.

--
Thomas T. Veldhouse
Key Fingerprint: 2DB9 813F F510 82C2 E1AE 34D0 D69D 1EDC D5EC AED1
--

In San Francisco, Halloween is redundant.
-- Will Durst

Date: 17 Mar 2006 05:37:47
From: Tony M
Subject: Re: Rnd2, CM9k vs Fritz 9k - 1 attachment

On 14 Mar 2006 16:39:54 GMT, "Thomas T. Veldhouse" <veldy71@yahoo.com >
wrote:

>Wilma <wilma_defemina@hotmail.com> wrote:
>> There is one item of relevant information you mention that I did not
>> include in my post or in the attachment heading: the hash tables. I used
>> the default settings that came with the programs. That's 37MB for Fritz9
>> and 4 MB for CM9k.
>>
>
>If you ran both engines against each other on the same machine, then you
>are not performing a valid test, especially if your machine does not
>have multiple CPUs. The results you are seeing may have less to do with
>the chess engines and more to do with the default configuration of each
>of them allowing one to grap more CPU time than the other.
>
>--
>Thomas T. Veldhouse
>Key Fingerprint: 2DB9 813F F510 82C2 E1AE 34D0 D69D 1EDC D5EC AED1

If both engines have pondering turned off, then they will not grab
each other's CPU time. Of course, with pondering off, the engines
will play at a reduced strength, which still may cause some to
question the validity of the test. Given that most folks today only
have a single computer with a single CPU, running tests with pondering
off is about the most accurate choice for those folks.

Tony

Date: 13 Mar 2006 02:12:19
From: Wilma
Subject: Re: Rnd2, CM9k vs Fritz 9k - 1 attachment

You didn't bother opening the attachment, did you?

Wilma

"George" <gtsavdar@auth.gr > wrote in news:1142212140.215579.228590
@e56g2000cwe.googlegroups.com:

>>Two rounds is enough for me. It's just too much trouble, especially when
> the results have been so clear.
>
> Obviously you don't have any idea about Chess-engines testing.....
>
> Please follow the following rules before you post another one:
> 1)Post the time controls you used for the match.
> 2)Post the hash table memory you gave to each program for the match.
> 3)Post the hardware the engines run.
> 4)Post the opening books each engine had.
> 5)Post the games(I see you gave an attachement but it appears as a
> complete mess to me).
> 6)Start making comments about the performace of each engine after 20
> matches and start making serious comments after 50 games. But even
> after 50 games you should not look very sure about the result. After
> 100 games you should be able to give with confidence a clear opinion
> about engine's strength. Even then, after 100 games, if the result is
> close enough, then you should play even more games....
>
>

Date: 13 Mar 2006 02:03:42
From: Wilma
Subject: Re: Rnd2, CM9k vs Fritz 9k - 1 attachment

No. :) Please check again. You didn't see what I posted because you were
too eager to lecture me. I'm done with it. You do a test and see what you
get. I'd honestly be interested in seeing what results you get.

Wilma

"George" <gtsavdar@auth.gr > wrote in news:1142212140.215579.228590
@e56g2000cwe.googlegroups.com:

>>Two rounds is enough for me. It's just too much trouble, especially when
> the results have been so clear.
>
> Obviously you don't have any idea about Chess-engines testing.....
>
> Please follow the following rules before you post another one:
> 1)Post the time controls you used for the match.
> 2)Post the hash table memory you gave to each program for the match.
> 3)Post the hardware the engines run.
> 4)Post the opening books each engine had.
> 5)Post the games(I see you gave an attachement but it appears as a
> complete mess to me).
> 6)Start making comments about the performace of each engine after 20
> matches and start making serious comments after 50 games. But even
> after 50 games you should not look very sure about the result. After
> 100 games you should be able to give with confidence a clear opinion
> about engine's strength. Even then, after 100 games, if the result is
> close enough, then you should play even more games....
>

Date: 13 Mar 2006 01:23:00
From: CeeBee
Subject: Re: Rnd2, CM9k vs Fritz 9k - 1 attachment

"George" <gtsavdar@auth.gr > wrote in rec.games.chess.computer:

> 5)Post the games(I see you gave an attachement but it appears as a
> complete mess to me).

(7) And don't post attachments in text groups.

--
CeeBee

*** The Cookie Has Spoken ***

Date: 13 Mar 2006 03:27:43
From: Wilma
Subject: Re: Rnd2, CM9k vs Fritz 9k - 1 attachment

CeeBee <ceebeechester@start.com.au > wrote in
news:Xns9785183EF4F90ceebeechesterstartco@213.75.12.164:

> "George" <gtsavdar@auth.gr> wrote in rec.games.chess.computer:
>
>> 5)Post the games(I see you gave an attachement but it appears as a
>> complete mess to me).
>
> (7) And don't post attachments in text groups.
>

Okay, here's the attachment in text form:

[Event "CM9k Fritz9 Match"]
[Site "Ram=255Mb 1.6Gh Run separately"]
[Date "2006.3.9"]
[Round "2"]
[White "Chessmaster 9000"]
[Black "Fritz9"]
[TimeControl "40/120"]
[Result "1-0"]
[ECO "C87"]

1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6 5.O-O d6 6.Re1 Be7 7.c3 O-O 8.d4
Bd7 9.Nbd2 exd4 10.cxd4 Nb4 11.Bb3 Nd3 12.Re3 Nxf2 13.Kxf2 Ng4+ 14.Ke2
Bb5+ 15.Nc4 f5 16.a4 Nxe3 17.Kxe3 Kh8 18.axb5 fxe4 19.Kxe4 d5+ 20.Ke3
dxc4 21.Bxc4 axb5 22.Rxa8 Qxa8 23.Bxb5 Qa5 24.Bd3 Qh5 25.Kd2 c5 26.dxc5
Qxc5 27.Ke2 Bd6 28.Qc2 Qb4 29.Kd1 Bf4 30.Bxh7 Bxc1 31.Kxc1 Rxf3 32.gxf3
Qe1+ 33.Qd1 Qe3+ 34.Kc2 Qc5+ 35.Kb1 Kxh7 36.Qc2+ Qxc2+ 37.Kxc2 Kg6
38.Kd3 1-0

Wilma

Date: 13 Mar 2006 02:07:46
From: Wilma
Subject: Re: Rnd2, CM9k vs Fritz 9k - 1 attachment

Don't make up rules for me. If I put the pgn in the text, someone
invariably suggests I post it as an attachment to make it easier to use. I
agree with that suggestion, not your peronal rule. Your post is tangential.

Wilma

CeeBee <ceebeechester@start.com.au > wrote in
news:Xns9785183EF4F90ceebeechesterstartco@213.75.12.164:

> "George" <gtsavdar@auth.gr> wrote in rec.games.chess.computer:
>
>> 5)Post the games(I see you gave an attachement but it appears as a
>> complete mess to me).
>
> (7) And don't post attachments in text groups.
>

Date: 13 Mar 2006 10:36:32
From: CeeBee
Subject: Re: Rnd2, CM9k vs Fritz 9k - 1 attachment

Wilma <wilma_defemina@hotmail.com > wrote in rec.games.chess.computer:

> Don't make up rules for me. If I put the pgn in the text, someone
> invariably suggests I post it as an attachment to make it easier to use.
> I agree with that suggestion, not your peronal rule.

I don't make up rules for you. Providers do. Most newsserver
automatically drop attachments or even the complete message with an
attachment in text groups.
Your original message was an example. It didn't show up on my newsserver
at all.

Furthermore not every newsreader is able to properly decode attachement you
encode when you post it as attachment, making itup a garbled mess; the
response of George to your message is an illustration of that.

Providers mostly restrict posts with
attachments either to the specific binaries groups or to their dedicated
binaries servers, where they limit rentention time on server level due to
space limitations.

> Your post is
> tangential.

Like brakes in a car are tangential when driving.

--
CeeBee

*** The Cookie Has Spoken ***

Date: 14 Mar 2006 07:33:00
From: Wilma
Subject: Re: Rnd2, CM9k vs Fritz 9k - 1 attachment

CeeBee <ceebeechester@start.com.au > wrote in
news:Xns978576183E37ceebeechesterstartco@213.75.12.164:

> Wilma <wilma_defemina@hotmail.com> wrote in rec.games.chess.computer:
>
>> Don't make up rules for me. If I put the pgn in the text, someone
>> invariably suggests I post it as an attachment to make it easier to
>> use. I agree with that suggestion, not your peronal rule.
>
> I don't make up rules for you. Providers do. Most newsserver
> automatically drop attachments or even the complete message with an

I stick to mainstream computer orgs like earthlink, so I tend to forget
that there are underserving organizations. There is no sucj geberal rule,
but what you say is true that not everyone can see attachments. Some want
them, however, and request them so they can see the game with having to
enter it move by move.
>
>
>> Your post is
>> tangential.
>
> Like brakes in a car are tangential when driving.
>

Narrow view of driving. Please stay off the road. :)

Wilma

Date: 14 Mar 2006 14:15:49
From: CeeBee
Subject: Re: Rnd2, CM9k vs Fritz 9k - 1 attachment

Wilma <wilma_defemina@hotmail.com > wrote in rec.games.chess.computer:

> I stick to mainstream computer orgs like earthlink, so I tend to forget
> that there are underserving organizations.

This has not much to do with "underserving" organisations, but, as I
explained to you before, with the difference between text based and groups
and binaries and with the accompanied server load and retention time.

> There is no sucj geberal rule,

It certainly is. _Properly_ configured news servers don't accept attachments
in text groups, only in binaries. The only difference is that some are kind
enough to drop the attachment only, while the standard behaviour is that
they drop the message altogether.

--
CeeBee

*** The Cookie Has Spoken ***

Date: 14 Mar 2006 12:53:21
From: David Richerby
Subject: Re: Rnd2, CM9k vs Fritz 9k - 1 attachment

Wilma <wilma_defemina@hotmail.com > wrote:
> CeeBee <ceebeechester@start.com.au> wrote:
>> Wilma <wilma_defemina@hotmail.com> wrote:
>>> Don't make up rules for me. If I put the pgn in the text, someone
>>> invariably suggests I post it as an attachment to make it easier
>>> to use. I agree with that suggestion, not your peronal rule.
>>
>> I don't make up rules for you. Providers do. Most newsserver
>> automatically drop attachments or even the complete message with an
>
> I stick to mainstream computer orgs like earthlink, so I tend to
> forget that there are underserving organizations.

Many organizations don't have the resources to store attachments to
news posts. There are enormous numbers of groups (tens of thousands)
and, given the choice of either holding a large attachment or several
tens of text posts, most news administrators will go for the text
posts. Admittedly, your attachment was small so would not have caused
problems but it's much easier to enforce `no attachments' than `only
small attachments.'

> There is no sucj geberal rule

Au contraire. Posting binaries (attachments) is against the AUP of
many ISPs. See also, for example,

http://www.uwasa.fi/~ts/http/nobin.html
http://www.cs.tut.fi/~jkorpela/usenet/dont.html

http://www.faqs.org/faqs/usenet/cancel-faq/part1/index.html

says,

``3. Third person cancels are generally frowned upon, unless they are
made based on one of the following criteria:

``e. Binaries in a non-binary group
Much of Usenet does not want binary messages, usually for disk
space and performance reasons. To accommodate those sites that
do want binaries, the alt.binaries.* and comp.binaries.*
hierarchies were created. However, there are still some binary
messages posted to other Usenet groups; these are often
cancelled without regard to content, based on the size of the
binary. For more information, see news.admin.net-abuse.policy,
where the specifics are being debated.''

> but what you say is true that not everyone can see Some want them,
> however, and request them so they can see the game with having to
> enter it move by move.

It's perfectly possible to post the moves of a game in the text of the
article, as you later did. I don't recall anyone ever asking for
anything to be posted as an attachment here, though it's entirely
possible that I've missed some or forgotten.

Dave.

--
David Richerby Devil Atom Bomb (TM): it's like a
www.chiark.greenend.org.uk/~davidr/ weapon of mass destruction that's
possessed by Satan!