Geschrieben von:/Posted by: Leo Dijksman at 22 September 2004 16:57:16:
Premier Division Division:
Round 8:
Shredder 7.04 0½1½ 2.0/4
Gandalf 4.32h 1½0½ 2.0/4
--------------------------------------
Tao 5.7b04 ½11½ 3.0/4
Nejmet 3.07 ½00½ 1.0/4
--------------------------------------
Aristarch 4.50 1½ 1.5/2 = 2 games to go!
Yace 0.99.87 0½ 0.5/2
--------------------------------------
There was a problem (blue screen) with my MP2400 monday, therefore
the Premier is moved to the MP2600!
2nd Division:
Round 10:
The Baron 1.4.0 1011 3.0/4
Betsy 6.51 0100 1.0/4
---------------------------------
Fruit X-0705 1111 4.0/4 !!
Ufim 5.01 0000 0.0/4
---------------------------------
KnightDreamer 3.3 1101 3.0/4 = lost one game on time! (in a bad position)
Terra 3.3b11 0010 1.0/4
---------------------------------
Jonny 2.64 ½1½ 2.0/3 = 1 game to go!
Chezzz 1.0.3 ½0½ 1.0/3
---------------------------------
4th Division:
Round 14:
GreKo 2.812 1110 3.0/4
Bodo 0.2e 0001 1.0/4
---------------------------------
Round 15:
ZZZZZZ 6.427 101½ 2.5/4
Celes 0.78 wbec7 010½ 1.5/4
---------------------------------
CyberPagno 2.0.1 1½0 1.5/3 = 1 game to go!
BlackBishop 0.9.7i 0½1 1.5/3
---------------------------------
Crosstables and pgn on the WBEC homepage.
Enginelist, latest updates/new engines:
22/09/2004:
Updated: Petir v1.0.1
Updated: GES v1.19
Updated: Ghost v0.14
Best wishes,
Leo.
WBEC Ridderkerk homepage.
Winboard Forum
Discussions (not only) about Xboard/Winboard, Computer Chess and Chess Programming
WBEC Ridderkerk new results.
10 posts
• Page 1 of 1
Re: WBEC Ridderkerk new results.
by Gábor Szots » 22 Sep 2004, 17:40
Geschrieben von:/Posted by: Gábor Szots at 22 September 2004 18:40:09:
Als Antwort auf:/In reply to: WBEC Ridderkerk new results. geschrieben von:/posted by: Leo Dijksman at 22 September 2004 16:57:16:
Hi Leo,
Something is wrong with the 2nd division crosstable. The rows and coloumns do not match, e.g. Naum is in a different row than coloumn.
Gábor
Als Antwort auf:/In reply to: WBEC Ridderkerk new results. geschrieben von:/posted by: Leo Dijksman at 22 September 2004 16:57:16:
Hi Leo,
Something is wrong with the 2nd division crosstable. The rows and coloumns do not match, e.g. Naum is in a different row than coloumn.
Gábor
- Gábor Szots
Re: WBEC Ridderkerk new results.
by Leo Dijksman » 22 Sep 2004, 19:48
Geschrieben von:/Posted by: Leo Dijksman at 22 September 2004 20:48:38:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von:/posted by: Gábor Szots at 22 September 2004 18:40:09:
I was forget to copy the top row, was still the old one from yesterday.
Is corrected now!
Best wishes,
Leo.
WBEC Ridderkerk homepage.
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von:/posted by: Gábor Szots at 22 September 2004 18:40:09:
Thanks Gábor!Hi Leo,
Something is wrong with the 2nd division crosstable. The rows and coloumns do not match, e.g. Naum is in a different row than coloumn.
Gábor
I was forget to copy the top row, was still the old one from yesterday.
Is corrected now!
Best wishes,
Leo.
WBEC Ridderkerk homepage.
- Leo Dijksman
Those WBTMs do not do 1 RR at a time
by Norm Pollock » 22 Sep 2004, 20:20
Geschrieben von:/Posted by: Norm Pollock at 22 September 2004 21:20:37:
Als Antwort auf:/In reply to: WBEC Ridderkerk new results. geschrieben von:/posted by: Leo Dijksman at 22 September 2004 16:57:16:
It is my understanding that the various winboard tournament managers do not play 1 RR at a time unless it is a 1 RR tournament. I hope that someday they will correct this situation.
As can be seen by the table of Div 2 in Leo's tournament, the number of games played varies widely. Frenzee has played 28 games while Fruit has played 40. Other engines are between these 2 extremes. The situation here has occurred because of 2 disqualifications and because of the way the wbtms do their job.
Even without disqualifications, the wbtm's tables have a lot of imbalance. In the Premier Division, the variation is 4. Some engines have played 28 while others have played 32.
But why so much imbalance? That is because the wbtm's play a 4RR tournament by matching each engine pairing 4 consecutive times. Instead it should do each pairing once before it does a pairing a second time. That way the table will show less variation of games played. The fritz8 gui does it that way.
Als Antwort auf:/In reply to: WBEC Ridderkerk new results. geschrieben von:/posted by: Leo Dijksman at 22 September 2004 16:57:16:
It is my understanding that the various winboard tournament managers do not play 1 RR at a time unless it is a 1 RR tournament. I hope that someday they will correct this situation.
As can be seen by the table of Div 2 in Leo's tournament, the number of games played varies widely. Frenzee has played 28 games while Fruit has played 40. Other engines are between these 2 extremes. The situation here has occurred because of 2 disqualifications and because of the way the wbtms do their job.
Even without disqualifications, the wbtm's tables have a lot of imbalance. In the Premier Division, the variation is 4. Some engines have played 28 while others have played 32.
But why so much imbalance? That is because the wbtm's play a 4RR tournament by matching each engine pairing 4 consecutive times. Instead it should do each pairing once before it does a pairing a second time. That way the table will show less variation of games played. The fritz8 gui does it that way.
- Norm Pollock
Re: Those WBTMs do not do 1 RR at a time
by Günther Simon » 22 Sep 2004, 20:32
Geschrieben von:/Posted by: Günther Simon at 22 September 2004 21:32:45:
Als Antwort auf:/In reply to: Those WBTMs do not do 1 RR at a time geschrieben von:/posted by: Norm Pollock at 22 September 2004 21:20:37:
should be played at once. The imbalance is just in the spectators eye,
what if Leo would release results less often, only to make you content?
BTW what about learning? All possible learning informations would be nearly
useless, if progs would be forced to split their matches.
If you want to do it that way you can do it easily with all WBTMs, just save
the first tourney file for RR1 and repeat it as often as you want.
(less than one minute work)
Günther
Als Antwort auf:/In reply to: Those WBTMs do not do 1 RR at a time geschrieben von:/posted by: Norm Pollock at 22 September 2004 21:20:37:
I disagree I would not like it otherways. If there are matches in RR theyIt is my understanding that the various winboard tournament managers do not play 1 RR at a time unless it is a 1 RR tournament. I hope that someday they will correct this situation.
As can be seen by the table of Div 2 in Leo's tournament, the number of games played varies widely. Frenzee has played 28 games while Fruit has played 40. Other engines are between these 2 extremes. The situation here has occurred because of 2 disqualifications and because of the way the wbtms do their job.
Even without disqualifications, the wbtm's tables have a lot of imbalance. In the Premier Division, the variation is 4. Some engines have played 28 while others have played 32.
But why so much imbalance? That is because the wbtm's play a 4RR tournament by matching each engine pairing 4 consecutive times. Instead it should do each pairing once before it does a pairing a second time. That way the table will show less variation of games played. The fritz8 gui does it that way.
should be played at once. The imbalance is just in the spectators eye,
what if Leo would release results less often, only to make you content?
BTW what about learning? All possible learning informations would be nearly
useless, if progs would be forced to split their matches.
If you want to do it that way you can do it easily with all WBTMs, just save
the first tourney file for RR1 and repeat it as often as you want.
(less than one minute work)
Günther
- Günther Simon
Re: Those WBTMs do not do 1 RR at a time
by Roger Brown » 23 Sep 2004, 01:57
Geschrieben von:/Posted by: Roger Brown at 23 September 2004 02:57:25:
Als Antwort auf:/In reply to: Re: Those WBTMs do not do 1 RR at a time geschrieben von:/posted by: Günther Simon at 22 September 2004 21:32:45:
Hello Norm,
Is this really critical to the sheer enjoyment that you get from Leo's tournament? I cannot imagine why this would be so....
Of course the Fritz gui is not a suitable replacement, excellent as it is.
Günther, do not even joke like that! Less often? Are you crazy? I cannot get enough as it is now! I believe that I am a fanatic. WBEC. RWBC. AEGT. Chesswar. Graham's long time control saga. More and more.
I can do some fun things with Jori's WBTM. The latest WBTM by Alex offers some options that I find fascinating.
Gunther, have you updated your main Thinker.dat book? The one I have is dated July 20, 2004 and is 2.5 Mb. Alex's WBTM uses the Thinker books so I am in the market for some good books. I have two (I think they are yours) at 14 and 30 plies at 45 and 253 Kb respectively.
Where are they to be found on your site incidentally?
Ahem, please remember my request for statistics on these books.
I think I have annoyed you enough for one day.
Later.
Als Antwort auf:/In reply to: Re: Those WBTMs do not do 1 RR at a time geschrieben von:/posted by: Günther Simon at 22 September 2004 21:32:45:
I disagree I would not like it otherways. If there are matches in RR theyBut why so much imbalance? That is because the wbtm's play a 4RR tournament by matching each engine pairing 4 consecutive times. Instead it should do each pairing once before it does a pairing a second time. That way the table will show less variation of games played. The fritz8 gui does it that way.
should be played at once. The imbalance is just in the spectators eye,
what if Leo would release results less often, only to make you content?
BTW what about learning? All possible learning informations would be nearly
useless, if progs would be forced to split their matches.
If you want to do it that way you can do it easily with all WBTMs, just save
the first tourney file for RR1 and repeat it as often as you want.
(less than one minute work)
Hello Norm,
Is this really critical to the sheer enjoyment that you get from Leo's tournament? I cannot imagine why this would be so....
Of course the Fritz gui is not a suitable replacement, excellent as it is.
Günther, do not even joke like that! Less often? Are you crazy? I cannot get enough as it is now! I believe that I am a fanatic. WBEC. RWBC. AEGT. Chesswar. Graham's long time control saga. More and more.
I can do some fun things with Jori's WBTM. The latest WBTM by Alex offers some options that I find fascinating.
Gunther, have you updated your main Thinker.dat book? The one I have is dated July 20, 2004 and is 2.5 Mb. Alex's WBTM uses the Thinker books so I am in the market for some good books. I have two (I think they are yours) at 14 and 30 plies at 45 and 253 Kb respectively.
Where are they to be found on your site incidentally?
Ahem, please remember my request for statistics on these books.
I think I have annoyed you enough for one day.
Later.
- Roger Brown
Re: Those WBTMs do not do 1 RR at a time
by Alex Guerrero » 23 Sep 2004, 04:07
Geschrieben von:/Posted by: Alex Guerrero at 23 September 2004 05:07:19:
Als Antwort auf:/In reply to: Those WBTMs do not do 1 RR at a time geschrieben von:/posted by: Norm Pollock at 22 September 2004 21:20:37:
Alex's WBTM does not matches in roundrobins tournaments, makes a round and after other like Fritz, although I will add the option for choosing the 2 modes.
Regards.
Als Antwort auf:/In reply to: Those WBTMs do not do 1 RR at a time geschrieben von:/posted by: Norm Pollock at 22 September 2004 21:20:37:
Hi Norman.It is my understanding that the various winboard tournament managers do not play 1 RR at a time unless it is a 1 RR tournament. I hope that someday they will correct this situation.
Alex's WBTM does not matches in roundrobins tournaments, makes a round and after other like Fritz, although I will add the option for choosing the 2 modes.
Regards.
- Alex Guerrero
Re: Those WBTMs do not do 1 RR at a time
by Norm Pollock » 23 Sep 2004, 05:09
Geschrieben von:/Posted by: Norm Pollock at 23 September 2004 06:09:43:
Als Antwort auf:/In reply to: Re: Those WBTMs do not do 1 RR at a time geschrieben von:/posted by: Günther Simon at 22 September 2004 21:32:45:
There are a couple of reasons why I prefer 1 at a time over 4 at a time in a 4 RR tournament:
(1) Balanced number of games finished. Example: Some engines have finished 32 games, while most others have finished 28 games, with 2 engines in between at 30. Wouldn't it be better for comparative purposes if all engines have finished 30 games?
(2) Balanced level of opponents. With 4 at a time, some engines have not played the top engines so their results are skewed to look better. (Perhaps use of S-B numbers would indicate their weaker schedule.) Likewise, some engines have faced the top engines but not the weak engines, so their results are skewed to look worse. In a 1 at a time system, at any time all engines have faced all other engines about the same number of times. The results therefore will have more meaning for comparative purposes.
--------------------
Btw, I don't know if book learning is better or worse with 1 or 4 at a time. Anyway I don't believe in book learning because it is dependent on the level of the opponent. Your engine can have a great opening but suppose the opponent is 200 elo higher, and your engine eventually loses. It is my understanding that book learning would mark that opening as a poor opening. Likewise, if facing a weaker opponent, and your engine wins despite a weak opening, book learning would mark that opening as a good opening.
Als Antwort auf:/In reply to: Re: Those WBTMs do not do 1 RR at a time geschrieben von:/posted by: Günther Simon at 22 September 2004 21:32:45:
There are a couple of reasons why I prefer 1 at a time over 4 at a time in a 4 RR tournament:
(1) Balanced number of games finished. Example: Some engines have finished 32 games, while most others have finished 28 games, with 2 engines in between at 30. Wouldn't it be better for comparative purposes if all engines have finished 30 games?
(2) Balanced level of opponents. With 4 at a time, some engines have not played the top engines so their results are skewed to look better. (Perhaps use of S-B numbers would indicate their weaker schedule.) Likewise, some engines have faced the top engines but not the weak engines, so their results are skewed to look worse. In a 1 at a time system, at any time all engines have faced all other engines about the same number of times. The results therefore will have more meaning for comparative purposes.
--------------------
Btw, I don't know if book learning is better or worse with 1 or 4 at a time. Anyway I don't believe in book learning because it is dependent on the level of the opponent. Your engine can have a great opening but suppose the opponent is 200 elo higher, and your engine eventually loses. It is my understanding that book learning would mark that opening as a poor opening. Likewise, if facing a weaker opponent, and your engine wins despite a weak opening, book learning would mark that opening as a good opening.
- Norm Pollock
Re: Those WBTMs do not do 1 RR at a time
by Norm Pollock » 23 Sep 2004, 05:22
Geschrieben von:/Posted by: Norm Pollock at 23 September 2004 06:22:27:
Als Antwort auf:/In reply to: Re: Those WBTMs do not do 1 RR at a time geschrieben von:/posted by: Roger Brown at 23 September 2004 02:57:25:
The color of the sides would always be the same.
-Norm
Als Antwort auf:/In reply to: Re: Those WBTMs do not do 1 RR at a time geschrieben von:/posted by: Roger Brown at 23 September 2004 02:57:25:
To use an analogy, wouldn't a sports league such as the English Premiership be a lot less interesting if teams played each other 2 at a time?Is this really critical to the sheer enjoyment that you get from Leo's >tournament? I cannot imagine why this would be so....
If you want to do it that way you can do it easily with all WBTMs, just save
the first tourney file for RR1 and repeat it as often as you want.
(less than one minute work)
The color of the sides would always be the same.
-Norm
- Norm Pollock
Re: Those WBTMs do not do 1 RR at a time
by Günther Simon » 24 Sep 2004, 16:23
Geschrieben von:/Posted by: Günther Simon at 24 September 2004 17:23:03:
Als Antwort auf:/In reply to: Re: Those WBTMs do not do 1 RR at a time geschrieben von:/posted by: Roger Brown at 23 September 2004 02:57:25:
my newest one. Just a little note: The 2.5MB one is not my main book,
but the book only for Shredder, TheKing and Fruit.
My main book is the 253KB one for UCI only progs without book
and the small one is in case 2 progs from above two groups
mentioned meet each other.
Well seems I have forgotten about this, or even worse, I know I had
done stats on this, but don't remember where I have saved the
results...
I don't feel bothered
Greets,
Günther
Als Antwort auf:/In reply to: Re: Those WBTMs do not do 1 RR at a time geschrieben von:/posted by: Roger Brown at 23 September 2004 02:57:25:
If you have those 3 you have all my books, the big one is also stillGunther, have you updated your main Thinker.dat book? The one I have is dated July 20, 2004 and is 2.5 Mb. Alex's WBTM uses the Thinker books so I am in the market for some good books. I have two (I think they are yours) at 14 and 30 plies at 45 and 253 Kb respectively.
Where are they to be found on your site incidentally?
Ahem, please remember my request for statistics on these books.
I think I have annoyed you enough for one day.
Later.
my newest one. Just a little note: The 2.5MB one is not my main book,
but the book only for Shredder, TheKing and Fruit.
My main book is the 253KB one for UCI only progs without book
and the small one is in case 2 progs from above two groups
mentioned meet each other.
Well seems I have forgotten about this, or even worse, I know I had
done stats on this, but don't remember where I have saved the
results...
I don't feel bothered
Greets,
Günther
- Günther Simon
10 posts
• Page 1 of 1
Return to Archive (Old Parsimony Forum)
Who is online
Users browsing this forum: No registered users and 69 guests