Winboard Forum

by **Leo Dijksman** » 14 Jan 2004, 17:16

Geschrieben von: / Posted by: Leo Dijksman at 14 January 2004 17:16:43:

New results Premier division:

Round 15:
Ruffian 2.0.0 11½½ 3.0/4
Dragon 4.5 00½½ 1.0/4
------------------------------------
Gandalf 4.32h 0111 3.0/4
Pharaon 2.62 1000 1.0/4
------------------------------------
GreenLightChess 3.00 1½ 1.5/2 = 2 games to go!
LittleGoliath 3.9po 0½ 0.5/2
------------------------------------

Next matches:
Round 15:
Aristarch 4.21 - Rebel 12
The King 3.23 - Crafty 19.06
WARP 0.58 - SmarThink 0.17a
Gromit 3.82 - Zarkov 4.5e
WBNimzo 2000b - Nejmet 3.07

New results 2nd Division:

Round 15:
King Of Kings 2.52 111½ 3.5/4
Butcher 1.42c 000½ 0.5/4
------------------------------------
Knightx 1.80a 101½ 2.5/4
Esc 1.16 010½ 1.5/4
------------------------------------
XiniX 2.9.11 0111 3.0/4
Amateur 2.70 1000 1.0/4
------------------------------------
Round 16:
Esc 1.16 - The Baron 1.1 b3 1-0 = 3 games to go!
------------------------------------

Next matches:
Round 16:
Amy 0.8.3po - Horizon 4.1
Chezzz 1.0.3 - Movei 0.08.131
Hagrid 0.7.56 - The CrazyBishop 0.04.5
List 4.61 - SlowChess 2.89b
Ktulu 4.1 - Resp 0.19
King Of Kings 2.52 - Queen 2.42
Butcher 1.42c - Amateur 2.70
XiniX 2.9.11 - Knightx 1.80a

Promo E tourney:
Round 2:
Gully 2.15 pl10 0½11 2.5/4
Cilian 4.13 1½00 1.5/4
------------------------------------
Round 3:
Gerbil 0.2 1111 4.0/4 !!
BigLion 2.23f 0000 0.0/4
------------------------------------
Sharper 0.17 0½11 2.5/4
Cilian 4.13 1½00 1.5/4
------------------------------------
Gully 2.15 pl10 1½0 1.5/3 = 1 game to go!
LaDameBlanche 2.0c 0½1 1.5/3
------------------------------------

Promo F tourney:
Round 3:
Uragano 0.43n 110½ 2.5/4
NoonianChess 3.6e 001½ 1.5/4
------------------------------------
Adam 1.6 1½10 2.5/4
SnailChess 4.013 0½01 1.5/4
------------------------------------
Round 4:
RDChess 3.23 1111 4.0/4 !!
Uragano 0.43n 0000 0.0/4
------------------------------------
SnailChess 4.013 ½101 2.5/4
Mustang 2.67 ½010 1.5/4
------------------------------------

Crosstables and pgn on the WBEC homepage.
Enginelist, latest updates/new engines:
14/01/2003:
Updated: Adamant v1.3 and v1.4
Leo.

WBEC Ridderkerk homepage.

by **Rudolf Posch** » 14 Jan 2004, 17:42

Geschrieben von: / Posted by: Rudolf Posch at 14 January 2004 17:42:04:
Als Antwort auf: / In reply to: WBEC Ridderkerk new results. geschrieben von: / posted by: Leo Dijksman at 14 January 2004 17:16:43:

....

Ugh, relief , I have been fearing already RDChess will have to play in the 6 th division.
But with the help of Uragano 0.43n it has not!
Rudolf

by **Matthias Gemuh** » 14 Jan 2004, 17:51

Geschrieben von: / Posted by: Matthias Gemuh at 14 January 2004 17:51:33:
Als Antwort auf: / In reply to: WBEC Ridderkerk new results. geschrieben von: / posted by: Leo Dijksman at 14 January 2004 17:16:43:

So functioning Pondering is absolutely necessary for BigLion !
I cannot say whether I shall fix it before Division 5 starts.
If not, BigLion will drop to Division 6 or lower.
/Matthias.

by **Uri Blass** » 14 Jan 2004, 18:00

Geschrieben von: / Posted by: Uri Blass at 14 January 2004 18:00:20:
Als Antwort auf: / In reply to: Re: WBEC Ridderkerk new results. geschrieben von: / posted by: Matthias Gemuh at 14 January 2004 17:51:33:

BigLion may also play in division 4 and promote to the premier.
The tournament is not over.
Pondering is not so important.
Little Goliath has no problem to play in the premier division without pondering.
Uri

by **Marcus Prewarski** » 14 Jan 2004, 18:11

Geschrieben von: / Posted by: Marcus Prewarski at 14 January 2004 18:11:58:
Als Antwort auf: / In reply to: Re: WBEC Ridderkerk new results. geschrieben von: / posted by: Uri Blass at 14 January 2004 18:00:20:

I think pondering is important especially in the lower divisions where dual pentiums are used instead of two separate computers.
Most people don't pay a lot of attention to lower divisions where many engines don't have pondering, but I've noticed a big difference in performance between these when you compare Leo's tournament to ones that don't use pondering.

-Marcus

by **Uri Blass** » 14 Jan 2004, 18:58

Geschrieben von: / Posted by: Uri Blass at 14 January 2004 18:58:54:
Als Antwort auf: / In reply to: Re: WBEC Ridderkerk new results. geschrieben von: / posted by: Marcus Prewarski at 14 January 2004 18:11:58:

I do not see what is the big difference between dual pentiums and seperate computers.

There is no big difference if the program does not have a special bugs in one of the cases.
difference in performance may be explained by other factors like different time control or different hardware(it is possible that one program likes P3 and another one prefers pentiun4).
Uri

by **Matthias Gemuh** » 14 Jan 2004, 20:49

Geschrieben von: / Posted by: Matthias Gemuh at 14 January 2004 20:49:28:
Als Antwort auf: / In reply to: Re: WBEC Ridderkerk new results. geschrieben von: / posted by: Uri Blass at 14 January 2004 18:00:20:

Hi Uri,
while BigLion was playing with Ponder=ON, he was topping his promotion group.
Then problems arose and Leo contacted me. I told him that BigLion should play
the rest of the tournament with Ponder=OFF. I then expected a real disaster,
and 0/4 is only the beginning of it.
I thought I had not modified protocol code since BigLion 2.23c, but I may be wrong.
Anyway, I shall surely have a more stable BigLion by the time he reaches Division 6.
/Matthias.

by **Albert Bertilsson** » 15 Jan 2004, 07:57

Geschrieben von: / Posted by: Albert Bertilsson at 15 January 2004 07:57:56:
Als Antwort auf: / In reply to: Re: WBEC Ridderkerk new results. geschrieben von: / posted by: Matthias Gemuh at 14 January 2004 17:51:33:

This was one of my first leassons in WBEC, without some efficient pondering you loose a lot of performance. It would be very sad to see Sharper promote because of a pondering problem in BigLion. It would feel much better if it managed to do so during normal conditions.
Well it ain't over yet, that 0.0/4 could just be an evil twist of bad luck, that happens so often when chess and statistics are involved.
/Regards Albert

by **Uri Blass** » 15 Jan 2004, 08:59

Geschrieben von: / Posted by: Uri Blass at 15 January 2004 08:59:20:
Als Antwort auf: / In reply to: Re: WBEC Ridderkerk new results. geschrieben von: / posted by: Albert Bertilsson at 15 January 2004 07:57:56:

I do not believe that you lose a lot of performance and it is probably a statistical illusion.
Little Goliath does not ponder and have no special problems in the premier division.
Pondering is overestimated
I believe that Shredder8 can beat every engine in 40/40 time control when Shredder does not ponder and the opponent ponders.
Uri

by **Matthias Gemuh** » 15 Jan 2004, 13:07

Geschrieben von: / Posted by: Matthias Gemuh at 15 January 2004 13:07:16:
Als Antwort auf: / In reply to: Re: WBEC Ridderkerk new results. geschrieben von: / posted by: Albert Bertilsson at 15 January 2004 07:57:56:

Hi Albert,
Uri is well-known for his strange views about pondering and engine strength .
Sharper is a worthy Division 4 engine, even without BigLion having problems,
so go for it !!!
/Matthias

by **Richard Pijl** » 15 Jan 2004, 14:14

Geschrieben von: / Posted by: Richard Pijl at 15 January 2004 14:14:25:
Als Antwort auf: / In reply to: Re: WBEC Ridderkerk new results. geschrieben von: / posted by: Uri Blass at 15 January 2004 08:59:20:

I tend to agree. Pondering will increase the thinking time the engine has for a position, probably about 50%. That's worthwhile implementing, but hardly a precondition for a strong engine. Engines that reuse the hashtable contents from previous iterations even suffer a bit at ponder misses as useful entries are probably being overwritten by useless ones.
Another advantage of a pondering engine when playing a non-pondering engine on a dual machine is that the pondering engine will have full access to all system resources (memory/cache, disk) while the non-pondering engine has not. I don't know how big that advantage is though.
Richard.

by **Marcus Prewarski** » 15 Jan 2004, 18:21

Geschrieben von: / Posted by: Marcus Prewarski at 15 January 2004 18:21:47:
Als Antwort auf: / In reply to: Re: WBEC Ridderkerk new results. geschrieben von: / posted by: Richard Pijl at 15 January 2004 14:14:25:

Yeah I'm a little curious about this myself. On a dual cpu system some resources are shared. It should be fairly simple to test. I notice that Leo has some crafty benchmarks for each of his systems, it should be as simple as running two benchmarks at the same time and comparing it with the single.
I'm a little less curious now that I've implemented pondering with my engine
-Marcus

Winboard Forum

Who is online