Super Tournament III Qualifying (standings after 22 of 68 ro

[Graham Banks]

Geschrieben von:/Posted by: Graham Banks at 21 July 2004 20:45:58:

Athlon 2000
256mb hash each (or closest allowed by each program)
3,4,5 piece tablebases
All using Fritz Powerbook tournament book (maximum variety, minimum 5 games for a move to be played, no learning)
40 moves in 40 minutes repeating.
Ponder off
Tournament run under Deep Fritz 8 GUI

STANDINGS (after 22 of 68 rounds)
14.5 - Pharaon 2.62
14.0 - List 512, Tao 5.6
13.5 - SOS 4
12.5 - Deep Sjeng 1.6, Delfi 1.6, Gandalf 5.1
12.0 - Little Goliath Revival, Thinker 4.6c
11.5 - Gothmog 1.0b7
11.0 - El Chinito 3.25
10.0 - Anaconda 1.6.2
9.5 - Yace 0.99.87
9.0 - Crafty 19.15
8.5 - Rebel 12, Comet B68
6.5 - Slow Chess 2.93a
6.0 - Fruit 1.5

[Roger Brown]

Geschrieben von:/Posted by: Roger Brown at 21 July 2004 21:10:11:
Als Antwort auf:/In reply to: Super Tournament III Qualifying (standings after 22 of 68 rounds) geschrieben von:/posted by: Graham Banks at 21 July 2004 20:45:58:

Athlon 2000
256mb hash each (or closest allowed by each program)
3,4,5 piece tablebases
All using Fritz Powerbook tournament book (maximum variety, minimum 5 games for a move to be played, no learning)
40 moves in 40 minutes repeating.
Ponder off
Tournament run under Deep Fritz 8 GUI

STANDINGS (after 22 of 68 rounds)
14.5 - Pharaon 2.62
14.0 - List 512, Tao 5.6
13.5 - SOS 4
12.5 - Deep Sjeng 1.6, Delfi 1.6, Gandalf 5.1
12.0 - Little Goliath Revival, Thinker 4.6c
11.5 - Gothmog 1.0b7
11.0 - El Chinito 3.25
10.0 - Anaconda 1.6.2
9.5 - Yace 0.99.87
9.0 - Crafty 19.15
8.5 - Rebel 12, Comet B68
6.5 - Slow Chess 2.93a
6.0 - Fruit 1.5

Hello Graham,
Your long timecontrol events are truly a labour of love and are appreciated...even if you do not get a lot of mail.

A few comments:
The order of these engines disturbs me. I know that you are less than half-way through so my concern may be premature but Crafty is languishing in the lower half of your table - behind Tao, Anaconda, Gothmog, Delfi and Thinker. Now before the authors of those engines have me for lunch I am not saying that their engines are tournament fodder. NO WAY.
I am saying that Crafty seems to be suffering under the gui you are using.
It cannot be incidental that the majority of engines in the topmost echelons support UCI directly. I recall you saying that the 1 mb bug did not afflict your setup. Does UCI send a new string for all the preceding moves or was that fixed? If no, then that would also explain some of it I guess. A WB engine suffers so.....
What is Delfi 1.6? That is not an engine I am familiar with.
Thanks much for the effort in running this everlasting tournament free for us all.
Later.

[Graham Banks]

Geschrieben von:/Posted by: Graham Banks at 21 July 2004 21:55:32:
Als Antwort auf:/In reply to: Re: Super Tournament III Qualifying (standings after 22 of 68 rounds) geschrieben von:/posted by: Roger Brown at 21 July 2004 21:10:11:

Athlon 2000
256mb hash each (or closest allowed by each program)
3,4,5 piece tablebases
All using Fritz Powerbook tournament book (maximum variety, minimum 5 games for a move to be played, no learning)
40 moves in 40 minutes repeating.
Ponder off
Tournament run under Deep Fritz 8 GUI

STANDINGS (after 22 of 68 rounds)
14.5 - Pharaon 2.62
14.0 - List 512, Tao 5.6
13.5 - SOS 4
12.5 - Deep Sjeng 1.6, Delfi 1.6, Gandalf 5.1
12.0 - Little Goliath Revival, Thinker 4.6c
11.5 - Gothmog 1.0b7
11.0 - El Chinito 3.25
10.0 - Anaconda 1.6.2
9.5 - Yace 0.99.87
9.0 - Crafty 19.15
8.5 - Rebel 12, Comet B68
6.5 - Slow Chess 2.93a
6.0 - Fruit 1.5

Hello Graham,
Your long timecontrol events are truly a labour of love and are appreciated...even if you do not get a lot of mail.

A few comments:
The order of these engines disturbs me. I know that you are less than half-way through so my concern may be premature but Crafty is languishing in the lower half of your table - behind Tao, Anaconda, Gothmog, Delfi and Thinker. Now before the authors of those engines have me for lunch I am not saying that their engines are tournament fodder. NO WAY.
I am saying that Crafty seems to be suffering under the gui you are using.
It cannot be incidental that the majority of engines in the topmost echelons support UCI directly. I recall you saying that the 1 mb bug did not afflict your setup. Does UCI send a new string for all the preceding moves or was that fixed? If no, then that would also explain some of it I guess. A WB engine suffers so.....
What is Delfi 1.6? That is not an engine I am familiar with.
Thanks much for the effort in running this everlasting tournament free for us all.
Later.

It's Delfi 4.5 - that's a typo on my part - sorry.
I am using the CB version of Crafty 19.15 - there is no lag and time usage seems more sensible than when I tried wcrafty using the wb2uci system.
Tao 5.6 is a strong well balanced program. It is the performance of Pharaon 2.62 that I think is the standout at present.
Crafty has NEVER performed well on my Athlon 2000.
And finally thanks for the thanks - much appreciated.
Graham.

[Kurt Utzinger]

Geschrieben von:/Posted by: Kurt Utzinger at 21 July 2004 22:32:31:
Als Antwort auf:/In reply to: Re: Super Tournament III Qualifying (standings after 22 of 68 rounds) geschrieben von:/posted by: Roger Brown at 21 July 2004 21:10:11:

[...]

I am saying that Crafty seems to be suffering under the gui you are using.
It cannot be incidental that the majority of engines in the topmost echelons support UCI directly. I recall you saying that the 1 mb bug did not afflict your setup. Does UCI send a new string for all the preceding moves or was that fixed? If no, then that would also explain some of it I guess. A WB engine suffers so.....

[...]

If your assumption is correct can perhaps be found out
in our AT 2004 (90m+30s) tourney which we are playing
with Wcrafty 19.15 and its own book under Arena-GUI
http://www.utzingerk.com/at_2004.htm
Kurt

[Joachim Rang]

Geschrieben von:/Posted by: Joachim Rang at 21 July 2004 22:59:33:
Als Antwort auf:/In reply to: Re: Super Tournament III Qualifying (standings after 22 of 68 rounds) geschrieben von:/posted by: Roger Brown at 21 July 2004 21:10:11:

Athlon 2000
256mb hash each (or closest allowed by each program)
3,4,5 piece tablebases
All using Fritz Powerbook tournament book (maximum variety, minimum 5 games for a move to be played, no learning)
40 moves in 40 minutes repeating.
Ponder off
Tournament run under Deep Fritz 8 GUI

STANDINGS (after 22 of 68 rounds)
14.5 - Pharaon 2.62
14.0 - List 512, Tao 5.6
13.5 - SOS 4
12.5 - Deep Sjeng 1.6, Delfi 1.6, Gandalf 5.1
12.0 - Little Goliath Revival, Thinker 4.6c
11.5 - Gothmog 1.0b7
11.0 - El Chinito 3.25
10.0 - Anaconda 1.6.2
9.5 - Yace 0.99.87
9.0 - Crafty 19.15
8.5 - Rebel 12, Comet B68
6.5 - Slow Chess 2.93a
6.0 - Fruit 1.5

Hello Graham,
Your long timecontrol events are truly a labour of love and are appreciated...even if you do not get a lot of mail.

A few comments:
The order of these engines disturbs me. I know that you are less than half-way through so my concern may be premature but Crafty is languishing in the lower half of your table - behind Tao, Anaconda, Gothmog, Delfi and Thinker. Now before the authors of those engines have me for lunch I am not saying that their engines are tournament fodder. NO WAY.
I am saying that Crafty seems to be suffering under the gui you are using.
It cannot be incidental that the majority of engines in the topmost echelons support UCI directly. I recall you saying that the 1 mb bug did not afflict your setup. Does UCI send a new string for all the preceding moves or was that fixed? If no, then that would also explain some of it I guess. A WB engine suffers so.....
What is Delfi 1.6? That is not an engine I am familiar with.
Thanks much for the effort in running this everlasting tournament free for us all.
Later.

well the Chessbasecrafty is dubious. First of all Graham did not deactivate the "asymmetric evaluation" which means Crafty thinks it plays against humans. Secondly there are other results which seem to indicate that crafty plays not at his full strength as a chessbase native.
regards Joachim

[Norm Pollock]

Geschrieben von:/Posted by: Norm Pollock at 21 July 2004 23:06:58:
Als Antwort auf:/In reply to: Re: Super Tournament III Qualifying (standings after 22 of 68 rounds) geschrieben von:/posted by: Graham Banks at 21 July 2004 21:55:32:

Athlon 2000
256mb hash each (or closest allowed by each program)
3,4,5 piece tablebases
All using Fritz Powerbook tournament book (maximum variety, minimum 5 games for a move to be played, no learning)
40 moves in 40 minutes repeating.
Ponder off
Tournament run under Deep Fritz 8 GUI

STANDINGS (after 22 of 68 rounds)
14.5 - Pharaon 2.62
14.0 - List 512, Tao 5.6
13.5 - SOS 4
12.5 - Deep Sjeng 1.6, Delfi 1.6, Gandalf 5.1
12.0 - Little Goliath Revival, Thinker 4.6c
11.5 - Gothmog 1.0b7
11.0 - El Chinito 3.25
10.0 - Anaconda 1.6.2
9.5 - Yace 0.99.87
9.0 - Crafty 19.15
8.5 - Rebel 12, Comet B68
6.5 - Slow Chess 2.93a
6.0 - Fruit 1.5

Hello Graham,
Your long timecontrol events are truly a labour of love and are appreciated...even if you do not get a lot of mail.

A few comments:
The order of these engines disturbs me. I know that you are less than half-way through so my concern may be premature but Crafty is languishing in the lower half of your table - behind Tao, Anaconda, Gothmog, Delfi and Thinker. Now before the authors of those engines have me for lunch I am not saying that their engines are tournament fodder. NO WAY.
I am saying that Crafty seems to be suffering under the gui you are using.
It cannot be incidental that the majority of engines in the topmost echelons support UCI directly. I recall you saying that the 1 mb bug did not afflict your setup. Does UCI send a new string for all the preceding moves or was that fixed? If no, then that would also explain some of it I guess. A WB engine suffers so.....
What is Delfi 1.6? That is not an engine I am familiar with.
Thanks much for the effort in running this everlasting tournament free for us all.
Later.

It's Delfi 4.5 - that's a typo on my part - sorry.
I am using the CB version of Crafty 19.15 - there is no lag and time usage seems more sensible than when I tried wcrafty using the wb2uci system.
Tao 5.6 is a strong well balanced program. It is the performance of Pharaon 2.62 that I think is the standout at present.
Crafty has NEVER performed well on my Athlon 2000.
And finally thanks for the thanks - much appreciated.
Graham.

Results are results. Conditions are the same for all the engines including the same book. Crafty is not that strong of an engine in this setting and with this level of competition.
I run Crafty in a very, very similar setting and my results concur very much with those above. However I found that using the wb version of Crafty (by BH with nosse) with the wb2uci adapter, is slightly stronger than the cb native version. I doubt the cb native version is compiled for the Athlon xp processors.

[Roger Brown]

Geschrieben von:/Posted by: Roger Brown at 21 July 2004 23:08:05:
Als Antwort auf:/In reply to: Re: Super Tournament III Qualifying (standings after 22 of 68 rounds) geschrieben von:/posted by: Kurt Utzinger at 21 July 2004 22:32:31:

If your assumption is correct can perhaps be found out
in our AT 2004 (90m+30s) tourney which we are playing
with Wcrafty 19.15 and its own book under Arena-GUI
http://www.utzingerk.com/at_2004.htm
Kurt

Hello Kurt,
I shall have you arrested for mental cruelty! How could you send me all excited to the site to gaze on an empty crosstable?
Oh the anguish, the pain.....
I await the games.


Later.

[Roger Brown]

Geschrieben von:/Posted by: Roger Brown at 21 July 2004 23:19:40:
Als Antwort auf:/In reply to: Re: Super Tournament III Qualifying (standings after 22 of 68 rounds) geschrieben von:/posted by: Norm Pollock at 21 July 2004 23:06:58:

Results are results. Conditions are the same for all the engines including the same book. Crafty is not that strong of an engine in this setting and with this level of competition.
I run Crafty in a very, very similar setting and my results concur very much with those above. However I found that using the wb version of Crafty (by BH with nosse) with the wb2uci adapter, is slightly stronger than the cb native version. I doubt the cb native version is compiled for the Athlon xp processors.

Hello Norm,
Results are not results in my opinion. Surely the results obtained have to be looked at in terms of the setup in effect?
For Crafty, the Winboard version, it is critical when playing against computers that the command computer (does arena send this command?) be inserted in the crafty rc or Crafty will assume that it is playing a human.

I do not know if the chessbase version has this switch but if it does not then the results have to be treated with some scepticism.

Note: not the person running the contest...just the results of the contest.

As you have also observed, the chessbase crafty may suffer under Athlons.

Surely those facts would have a significant effect on the outcome of a match?



Later.

[Norm Pollock]

Geschrieben von:/Posted by: Norm Pollock at 21 July 2004 23:42:41:
Als Antwort auf:/In reply to: Re: Super Tournament III Qualifying (standings after 22 of 68 rounds) geschrieben von:/posted by: Roger Brown at 21 July 2004 23:19:40:

Results are results. Conditions are the same for all the engines including the same book. Crafty is not that strong of an engine in this setting and with this level of competition.
I run Crafty in a very, very similar setting and my results concur very much with those above. However I found that using the wb version of Crafty (by BH with nosse) with the wb2uci adapter, is slightly stronger than the cb native version. I doubt the cb native version is compiled for the Athlon xp processors.

Hello Norm,
Results are not results in my opinion. Surely the results obtained have to be looked at in terms of the setup in effect?
For Crafty, the Winboard version, it is critical when playing against computers that the command computer (does arena send this command?) be inserted in the crafty rc or Crafty will assume that it is playing a human.

I do not know if the chessbase version has this switch but if it does not then the results have to be treated with some scepticism.

Note: not the person running the contest...just the results of the contest.

As you have also observed, the chessbase crafty may suffer under Athlons.

Surely those facts would have a significant effect on the outcome of a match?



Later.

If you use wb2uci with a wb version of Crafty, the crafty.rc file is read. In crafty.rc I have the computer command. I also have learning 7. What is interesting is that even though I have learning "off" in the fritz gui, the position.lrn and book.lrn files get updated but the fritz gui book does not. Since I get similar results, I don't think that the "computer" command makes any difference.
It would prefer that Crafty be UCI as well. But that is not my decision. That is completely up to Prof. Hyatt.

[Graham Banks]

Geschrieben von:/Posted by: Graham Banks at 22 July 2004 00:26:35:
Als Antwort auf:/In reply to: Re: Super Tournament III Qualifying (standings after 22 of 68 rounds) geschrieben von:/posted by: Joachim Rang at 21 July 2004 22:59:33:

Athlon 2000
256mb hash each (or closest allowed by each program)
3,4,5 piece tablebases
All using Fritz Powerbook tournament book (maximum variety, minimum 5 games for a move to be played, no learning)
40 moves in 40 minutes repeating.
Ponder off
Tournament run under Deep Fritz 8 GUI

STANDINGS (after 22 of 68 rounds)
14.5 - Pharaon 2.62
14.0 - List 512, Tao 5.6
13.5 - SOS 4
12.5 - Deep Sjeng 1.6, Delfi 1.6, Gandalf 5.1
12.0 - Little Goliath Revival, Thinker 4.6c
11.5 - Gothmog 1.0b7
11.0 - El Chinito 3.25
10.0 - Anaconda 1.6.2
9.5 - Yace 0.99.87
9.0 - Crafty 19.15
8.5 - Rebel 12, Comet B68
6.5 - Slow Chess 2.93a
6.0 - Fruit 1.5

Hello Graham,
Your long timecontrol events are truly a labour of love and are appreciated...even if you do not get a lot of mail.

A few comments:
The order of these engines disturbs me. I know that you are less than half-way through so my concern may be premature but Crafty is languishing in the lower half of your table - behind Tao, Anaconda, Gothmog, Delfi and Thinker. Now before the authors of those engines have me for lunch I am not saying that their engines are tournament fodder. NO WAY.
I am saying that Crafty seems to be suffering under the gui you are using.
It cannot be incidental that the majority of engines in the topmost echelons support UCI directly. I recall you saying that the 1 mb bug did not afflict your setup. Does UCI send a new string for all the preceding moves or was that fixed? If no, then that would also explain some of it I guess. A WB engine suffers so.....
What is Delfi 1.6? That is not an engine I am familiar with.
Thanks much for the effort in running this everlasting tournament free for us all.
Later.

well the Chessbasecrafty is dubious. First of all Graham did not deactivate the "asymmetric evaluation" which means Crafty thinks it plays against humans. Secondly there are other results which seem to indicate that crafty plays not at his full strength as a chessbase native.
regards Joachim

Remember that I used wcrafty with the wb2uci adaptor and it had scored 7/20 before I replaced it with CB Crafty and reran all the games.
Graham.

[Graham Banks]

Geschrieben von:/Posted by: Graham Banks at 22 July 2004 00:30:33:
Als Antwort auf:/In reply to: Re: Super Tournament III Qualifying (standings after 22 of 68 rounds) geschrieben von:/posted by: Norm Pollock at 21 July 2004 23:42:41:

Results are results. Conditions are the same for all the engines including the same book. Crafty is not that strong of an engine in this setting and with this level of competition.
I run Crafty in a very, very similar setting and my results concur very much with those above. However I found that using the wb version of Crafty (by BH with nosse) with the wb2uci adapter, is slightly stronger than the cb native version. I doubt the cb native version is compiled for the Athlon xp processors.

Hello Norm,
Results are not results in my opinion. Surely the results obtained have to be looked at in terms of the setup in effect?
For Crafty, the Winboard version, it is critical when playing against computers that the command computer (does arena send this command?) be inserted in the crafty rc or Crafty will assume that it is playing a human.

I do not know if the chessbase version has this switch but if it does not then the results have to be treated with some scepticism.

Note: not the person running the contest...just the results of the contest.

As you have also observed, the chessbase crafty may suffer under Athlons.

Surely those facts would have a significant effect on the outcome of a match?



Later.

If you use wb2uci with a wb version of Crafty, the crafty.rc file is read. In crafty.rc I have the computer command. I also have learning 7. What is interesting is that even though I have learning "off" in the fritz gui, the position.lrn and book.lrn files get updated but the fritz gui book does not. Since I get similar results, I don't think that the "computer" command makes any difference.
It would prefer that Crafty be UCI as well. But that is not my decision. That is completely up to Prof. Hyatt.

However disappointing that may be!
Graham.

[Roger Brown]

Geschrieben von:/Posted by: Roger Brown at 22 July 2004 02:09:26:
Als Antwort auf:/In reply to: You may have to accept that Crafty isn't as strong as you'd hoped! geschrieben von:/posted by: Graham Banks at 22 July 2004 00:30:33:

However disappointing that may be!
Graham.

Hello Graham,
I will go wherever the facts may lead. However before I am prepared to accept the veracity of your banner headline I have to be convinced of the methodology used to arrive at the conclusions posted.
We have a version of Crafty compiled for a specific gui. No-one except Chessbase knows what that compiled engine contains.
We have the fact that the CB version does not work well on your machine anyway.
We have the Winboard executable (assuming CB gui) being saddled with wb2UCI and an interface that sends instructions in a way that harms the engine's performance.
Finally, we have the fact that in long timecontrol events in winboard, Crafty performs somewhat better than the engines currently ahead of it.
With all of that, would I be prepared to say that Crafty was not as strong as engine X? I think not.
Your mileage may vary of course.
Looking forward to more results as your contest goes along....
Later.

[Thomas Mayer]

Geschrieben von:/Posted by: Thomas Mayer at 22 July 2004 02:55:02:
Als Antwort auf:/In reply to: Re: You may have to accept that Crafty isn't as strong as you'd hoped! geschrieben von:/posted by: Roger Brown at 22 July 2004 02:09:26:

Hi Roger,

We have a version of Crafty compiled for a specific gui. No-one except
Chessbase knows what that compiled engine contains.

exactly THAT is the point: Internally Crafty has the logic of a winboard-engine. Somehow (and noone knows how) ChessBase adjusts that to the needs of their GUI - as far as I know their protocol logic is more or less similar to UCI. I do not know how much effort and time they have spent initially in the ChessBase native version, but Mathias told me once that he simply exchange the interface routines and no other parts of the engine itself. It is unlikely that this has no impact to the playing strength at all. A proof to that is, that several testers have reported strength differences between the natives and the wb version - of several Crafty versions... -> how huge this difference is is unclear, I think nearly every number between 1 and 100 is possible.
Anyway it's definitely best to test with the WB-Crafty and not with the native because we really do not know what it exactly is...
Greets, Thomas

[Graham Banks]

Geschrieben von:/Posted by: Graham Banks at 22 July 2004 03:55:26:
Als Antwort auf:/In reply to: Re: You may have to accept that Crafty isn't as strong as you'd hoped! geschrieben von:/posted by: Thomas Mayer at 22 July 2004 02:55:02:

Apart from the 'computer' command, I ran wcrafty (AMD compile from Peter Skinner's site) as it should with the latest wb2uci adapter and a correctly configured crafty.rc file.
For the CB Crafty I've always used the default engine parameters.
Results have always been poor for Crafty on my Athlon no matter what (and I have run a lot of tournaments!), yet a friend of mine with a dual Athlon using ponder on has always had strong results using exactly the same CB Crafty as I have.
It is very difficult to draw concrete conclusions from these facts. One post directed those interested to an experiment that Volker had done to compare ponder off with ponder on and the outcome was that it didn't have much bearing at all on relative results of programs.
I do find it interesting though that some of you have had similar results to myself. Haven't had too much evidence in the way of tournament crosstables that Crafty 19.15 has performed exceptionally well for some members.
Anyway, on with the tournament and hopefully Crafty might be able to claw its way up the table!
Cheers, Graham.

[Heinz van Kempen]

Geschrieben von:/Posted by: Heinz van Kempen at 22 July 2004 06:37:05:
Als Antwort auf:/In reply to: Super Tournament III Qualifying (standings after 22 of 68 rounds) geschrieben von:/posted by: Graham Banks at 21 July 2004 20:45:58:

Athlon 2000
256mb hash each (or closest allowed by each program)
3,4,5 piece tablebases
All using Fritz Powerbook tournament book (maximum variety, minimum 5 games for a move to be played, no learning)
40 moves in 40 minutes repeating.
Ponder off
Tournament run under Deep Fritz 8 GUI

STANDINGS (after 22 of 68 rounds)
14.5 - Pharaon 2.62
14.0 - List 512, Tao 5.6
13.5 - SOS 4
12.5 - Deep Sjeng 1.6, Delfi 1.6, Gandalf 5.1
12.0 - Little Goliath Revival, Thinker 4.6c
11.5 - Gothmog 1.0b7
11.0 - El Chinito 3.25
10.0 - Anaconda 1.6.2
9.5 - Yace 0.99.87
9.0 - Crafty 19.15
8.5 - Rebel 12, Comet B68
6.5 - Slow Chess 2.93a
6.0 - Fruit 1.5

Hello Graham,
well, I am also playing with Crafty under Fritz GUI (but 19.14 version) and to be more accurate with a Bryan Hofmann compile especially for Athlon single CPU. I am enthusiastic with performance and with last Nunn position 13 I still had not the time to upload it is even better. The difference is that you are running games with 40/40 (what is fine and what we will also do in our AEGT) and I ran more than 600 Blitz games so far. So there is no comparison possible.
I only want to repeat over and over again...independent of time control 22, 50, 100, 200, even 400 games are not enough to be not prone to all sorts of statistical anomalies for results.
The new Crafties are superstrong (my opinion). Thanks for your fantastic tournament.
Best Regards
Heinz

[Bryan Hofmann]

Geschrieben von:/Posted by: Bryan Hofmann at 22 July 2004 15:26:50:
Als Antwort auf:/In reply to: My last comment on Crafty. geschrieben von:/posted by: Graham Banks at 22 July 2004 03:55:26:

Apart from the 'computer' command, I ran wcrafty (AMD compile from Peter Skinner's site) as it should with the latest wb2uci adapter and a correctly configured crafty.rc file.
For the CB Crafty I've always used the default engine parameters.
Results have always been poor for Crafty on my Athlon no matter what (and I have run a lot of tournaments!), yet a friend of mine with a dual Athlon using ponder on has always had strong results using exactly the same CB Crafty as I have.
It is very difficult to draw concrete conclusions from these facts. One post directed those interested to an experiment that Volker had done to compare ponder off with ponder on and the outcome was that it didn't have much bearing at all on relative results of programs.
I do find it interesting though that some of you have had similar results to myself. Haven't had too much evidence in the way of tournament crosstables that Crafty 19.15 has performed exceptionally well for some members.
Anyway, on with the tournament and hopefully Crafty might be able to claw its way up the table!
Cheers, Graham.

I can say that reports in the Crafty newsgroup have my AMD compiles up to 15% faster than Skinners.

On the CCC forum I have seen reports of using Crafty under the Fritz GUI and the problem is that the GUI sends the NEW command before every more. This has a big impact on Crafty as the NEW command tells Crafty it is starting a new game and to wipe out all hashing. With pondering off this effect is even more pronounced.