Computerchess Misc (8)
Rolf Tueschen
Before I want to write the longer article about computerchess testing
standards let me give you a simple example for my theory that a famous
tester on rgcc does something strange to weaken FRITZ. No wonder because
he always wanted to prove that FRITZ like ChessBase must be bad. Only
because he has some axe to grind with officials of ChessBase.
Parameters:
° processor type (it's known that FRITZ is designed for INTEL PENTIUM)
° memory (it's known that FRITZ runs best at >64 Mb; I dont know what
type is the best, SDRAM or what...)
° book/ settings
° engine/ settings (FRITZ plays his best chess in rated mode)
Note that we have for the "summer-tournament" e.g. no evidence for any
special setting, for the consistance of the settings as such. All what
we could do is trying to replay the game scores with our machines.
When I looked at the games of FRITZ where he played horrible chess I
discovered that my version of FRITZ5 preferred totally different moves
while the allegedly original moves were not even considered among the
first 5 alternatives.
But enough words, here're the two games. I melted the scores with CB7.
COMP JUNIOR5 - COMP FRITZ5 [A77]
summer-tournament (5), 23.08.1998
1.c4 ... before we continue here comes the other game:
[1.e4 e5 2.Nf3 Nc6 3.Bb5 Bc5 4.c3 Nge7 5.0-0 Bb6 6.d4 exd4 7.cxd4 d5
8.exd5 Nxd5 9.Re1+ Be6 10.Ne5 Nde7
now this position is the crucial one:
11.Bg5 ??
not reproduceble! (and for good reasons because this line is bad)
I have 4 variations in the following order from best to forth:
(Fritz 5.00: 11.Nf3 Qd6 12.Bg5 Kf8 13.Bxe7+ Nxe7 14.Nc3 Bg4 15.Re3 c6;
Fritz 5.00: 11.Bxc6+ Nxc6 12.Nxc6 bxc6 13.Be3 0-0 14.Nc3 Qd7 15.Rc1
Rad8;
Fritz 5.00: 11.Be3 0-0 12.Nxc6 Nxc6 13.Bxc6 bxc6;
Fritz 5.00: 11.Re4 0-0 12.Bxc6 Nxc6 13.Nxc6 bxc6 14.Nc3 Re8 15.Be3 Qd7)
11...Qxd4 12.Qxd4 Bxd4 13.Bxe7 0-1 COMP FRITZ5-COMP PHALANX/
1998/INTERNET CC98 (53)]
now back to the first game:
1...Nf6 2.d4 e6 3.Nf3 c5 4.d5 exd5 5.cxd5 d6 6.Nc3 g6 7.Nd2 Nbd7 8.e4
Bg7 9.Be2 0-0 10.0-0 Re8 11.a4 Ne5 12.Ra3 g5 13.Re1 a6 14.Nf1 h6 15.Ne3
Ng6 16.Qc2N Nf4 17.Bf3 b6 18.Nc4 N4xd5 19.Nxd5 Nxd5 20.Qd1 Nb4 21.Nxd6
Re7 22.e5 Rb8 23.Rae3 Be6 24.b3 Qc7 25.Bb2 Rd7 26.Qb1 Rbd8 27.Be4 Nd5
28.Rf3 Nf4 29.Bh7+ Kh8 30.Bf5
the crucial position again:
30... Qa7 ????
"no way!" This move doesn't make sense it loses without defense. More,
it is completely losing grip on the game. Nowhere that FRITZ should play
Qa7. What about a little manipulation of the drw-prefix? Just a thought.
*My* FRITZ5 plays
30...Qc6! 31.h4 +/= jun46 31...Qd5 (31...Nxg2 32.Be4 Nxe1 33.Bxc6 Nxf3+
34.Bxf3 gxh4 +/=) 32.hxg5 +- j46 32...hxg5 33.Be4 Qd2 +/= crafty 34.Bc3
Ne2+ 35.Rxe2 Qxe2 36.Bd3 Qxf3 37.gxf3 Rxd6 38.exd6 Bxc3 39.Bxa6 Rxd6
40.Bc4 Bxc4 41.bxc4 and I would prefer Be5!
"black defense can not be broken" RWT
(41...Bd2? 42.a5!) ]
31.h4 Qc7 32.hxg5 hxg5 33.Bc1 Bxf5 34.Qxf5 Re7 35.Bxf4 Rxd6 36.exd6
Rxe1+ 37.Kh2 Qd8 38.Qxf7 ... 1-0
Any comments?
Rolf Tueschen
<75mpm6$lre$1...@news05.btx.dtag.de>:
>[all snipped]
I forgot to ask you for three missing game-scores out of the quoted
summer-tournament:
a) Comet-Virtual
b) Diep-ChessM
c) Diogenes-Socrates
all games of #6.
Thank you.
mclane
>(promise for some older stuff; I will come back soon)
firstival rolf i will tell you why nobody answers your chess stuff:
because YOU have made this place into a NON-chess-place by spamming it
for years.
so if nobody wants to talk with you about the summer tournament, that
could be the reason. YOU are the reason :-)))
second:
you overestimate the amount of people who replay chess-games in this
newsgroup. this newsgroup is not about computerchess, it is about rolf
is calling us nazi's or whatever...:-))
> let me give you a simple example for my theory that a famous
>tester on rgcc does something strange to weaken FRITZ.
i don't have to WEAKEN fritz. it is weak enough.
> No wonder because
>he always wanted to prove that FRITZ like ChessBase must be bad. Only
>because he has some axe to grind with officials of ChessBase.
>Parameters:
What is this ?
parameters ?
parameters is that you don't have the same version of fritz i had. you
don't even try to find out about, which version i used, because in the
game you quoted
there is no real name of the program version that played the game.
you name it :
COMP JUNIOR5 - COMP FRITZ5 [A77]
I don't know the source you got it from. i can only guess. but WHAT i
know is, that this is not the exact version definition that i
published !
so it looks that you steal my copyrights :-)))) by renaming the games
and players !! :-))
whatever.
this way you will never be able to replay the games accurate, because:
you don't know which versions have played.
it was NOT junior , it was junior5 beta 16 bit !
a program you don't have , right ?
and it was not fritz5 (whatever fritz this might be!).
in my original postings of the summer tournament the original names of
the programs were published.
the source you quote is wrong. your results are wrong.
wrong input, wrong mind, wrong output.
>° processor type (it's known that FRITZ is designed for INTEL PENTIUM)
it's known that AMD cpu's are the best for several programs alike
rebel, ... I am not interested in what fritz is designed for. they get
the same hardware. and if fritz is so stupid that it runs slower on
this hardware, this seems to be a construction problem of the company,
not mine.
>° memory (it's known that FRITZ runs best at >64 Mb; I dont know what
>type is the best, SDRAM or what...)
fritz gets the same amount of memory the others get.
>° book/ settings
exactly. this would mean: you will have to use the power-books.
>° engine/ settings (FRITZ plays his best chess in rated mode)
very accurate rolf. fritz plays his best chess when they cheat, in
engine-engine fights, where they have presented us over months faked
results, or in autoplayer-games, where they constructed a cheating
autoplayer-device to get better results.
this is where fritz gets good results.
>Note that we have for the "summer-tournament" e.g. no evidence for any
>special setting, for the consistance of the settings as such.
you have NO evidence. and you will never have any evidence. because i
doubt that you are normal person rolf. you are ill. ill persons have
no evidence that they exist.
> All what
>we could do is trying to replay the game scores with our machines.
and you should make sure you replay them in the same conditions. if
you have different conditions, you cannot replay them. it's easy,
isn't it ?
so the first question would be: am i using the same machine.
am i using the same version.
do you ?
>When I looked at the games of FRITZ where he played horrible chess
we know this.
> I
>discovered that my version of FRITZ5 preferred totally different moves
exactly. MY version also always plays weaker moves than the
success-games chessBase publishes.
>while the allegedly original moves were not even considered among the
>first 5 alternatives.
aha. it seems your settings are different.
>But enough words, here're the two games. I melted the scores with CB7.
oh god !
you melted the score with cb7 ! unbelievable !
>COMP JUNIOR5 - COMP FRITZ5 [A77]
>summer-tournament (5), 23.08.1998
>1.c4 ... before we continue here comes the other game:
>[1.e4 e5 2.Nf3 Nc6 3.Bb5 Bc5 4.c3 Nge7 5.0-0 Bb6 6.d4 exd4 7.cxd4 d5
>8.exd5 Nxd5 9.Re1+ Be6 10.Ne5 Nde7
>now this position is the crucial one:
>11.Bg5 ??
?!?! which game, and what another game ?! can't you come with ONE game
after the other :-))))
multiple personalities ?! :-)))
>not reproduceble! (and for good reasons because this line is bad)
Bg5 is not reproducable ? and you replayed it with what ???
do you have the beta version amir sent me ?
ah - of course you have CHECKED anything. i forgot you have melted
anything with cb7, whatever this means.
>I have 4 variations in the following order from best to forth:
>(Fritz 5.00: 11.Nf3 Qd6 12.Bg5 Kf8 13.Bxe7+ Nxe7 14.Nc3 Bg4 15.Re3 c6;
>Fritz 5.00: 11.Bxc6+ Nxc6 12.Nxc6 bxc6 13.Be3 0-0 14.Nc3 Qd7 15.Rc1
>Rad8;
>Fritz 5.00: 11.Be3 0-0 12.Nxc6 Nxc6 13.Bxc6 bxc6;
>Fritz 5.00: 11.Re4 0-0 12.Bxc6 Nxc6 13.Nxc6 bxc6 14.Nc3 Re8 15.Be3 Qd7)
>11...Qxd4 12.Qxd4 Bxd4 13.Bxe7 0-1 COMP FRITZ5-COMP PHALANX/
>1998/INTERNET CC98 (53)]
the question remains: which fritz do you use ?
>now back to the first game:
puha !
>1...Nf6 2.d4 e6 3.Nf3 c5 4.d5 exd5 5.cxd5 d6 6.Nc3 g6 7.Nd2 Nbd7 8.e4
>Bg7 9.Be2 0-0 10.0-0 Re8 11.a4 Ne5 12.Ra3 g5 13.Re1 a6 14.Nf1 h6 15.Ne3
>Ng6 16.Qc2N Nf4 17.Bf3 b6 18.Nc4 N4xd5 19.Nxd5 Nxd5 20.Qd1 Nb4 21.Nxd6
>Re7 22.e5 Rb8 23.Rae3 Be6 24.b3 Qc7 25.Bb2 Rd7 26.Qb1 Rbd8 27.Be4 Nd5
>28.Rf3 Nf4 29.Bh7+ Kh8 30.Bf5
>the crucial position again:
>30... Qa7 ????
>"no way!" This move doesn't make sense it loses without defense. More,
>it is completely losing grip on the game. Nowhere that FRITZ should play
>Qa7. What about a little manipulation of the drw-prefix? Just a thought.
of course i have manipulated this move. :-)))
to be more precise: ossi weiner paid me 100000 $ to make fritz lose,
or was it amir ? i don't remember it anymore accurate enough :-)))
yes - it was amir. and he said he would like to see fritz lose. and he
promised that he would sign with a disclaimer next time :-)))
>*My* FRITZ5 plays
but what is YOUR fritz ?
>Any comments?
which machines do you use. which versions.
could you please come with ONE example after the other. it is very
difficult to read otherwise. if you quote games, please name the right
versions. i have published the data with accurate versions. why do YOU
now mix it up with noname FRITZ5 ? Which fritz5 ?
when i tell you i have a volkswagen, how do you know WHICH volkswagen
i do have ?
or is a golf the same than a lupo or a passat is ?
doesn't matter anyway ???
best wishes