Yace against Phalanx.

[Mogens Larsen]

Geschrieben von: / Posted by: Mogens Larsen at 22 June 2000 11:15:30:
I ran a 20 game blitz match (40/5') between Phalanx and Yace. The hardware was an Athlon 800MHz and the hashsize was 32Mb. Ponder off goes without saying.
The result was a crushing win by Yace:
Yace 0.18 - Phalanx 22C: 13.5-6.5.
Best wishes...
Mogens

[Aaron]

Geschrieben von: / Posted by: Aaron at 25 June 2000 16:14:58:
Als Antwort auf: / As an answer to: Yace against Phalanx. geschrieben von: / posted by: Mogens Larsen at 22 June 2000 11:15:30:

I ran a 20 game blitz match (40/5') between Phalanx and Yace. The hardware was an Athlon 800MHz and the hashsize was 32Mb. Ponder off goes without saying.
The result was a crushing win by Yace:
Yace 0.18 - Phalanx 22C: 13.5-6.5.
Best wishes...
Mogens

Are you convinced now that Yace is one of the top ranked winboard engines at Blitz ? Only Crafty and goliath maybe can match it..

[Mogens Larsen]

Geschrieben von: / Posted by: Mogens Larsen at 25 June 2000 16:23:40:
Als Antwort auf: / As an answer to: Re: Yace against Phalanx. geschrieben von: / posted by: Aaron at 25 June 2000 16:14:58:

Are you convinced now that Yace is one of the top ranked winboard engines at Blitz ? Only Crafty and goliath maybe can match it..

I always thought it was _one_ of the top ranked WinBoard engines at blitz. I just wasn't convinced about whether it was top ten or top five.
A few minor tests suggest that TCB and Yace are very close when it comes to blitz (6-5 to TCB at 40/5'), so I'll run a test sometime soon to clarify things further.
I believe that Crafty, Lg2000, AnMon, Yace and TCB are the best blitz programs, but that's only an opinion.
Best wishes...
Mogens

[Aaron]

Geschrieben von: / Posted by: Aaron at 25 June 2000 18:29:49:
Als Antwort auf: / As an answer to: Re: Yace against Phalanx. geschrieben von: / posted by: Mogens Larsen at 25 June 2000 16:23:40:

Are you convinced now that Yace is one of the top ranked winboard engines at >>Blitz ? Only Crafty and goliath maybe can match it..

I always thought it was _one_ of the top ranked WinBoard engines at blitz. I >just wasn't convinced about whether it was top ten or top five.
A few minor tests suggest that TCB and Yace are very close when it comes to >blitz (6-5 to TCB at 40/5'), so I'll run a test sometime soon to clarify >things further.
I believe that Crafty, Lg2000, AnMon, Yace and TCB are the best blitz >programs, but that's only an opinion.

Okay different definition of "best". For you best doesn't mean top 10?

I see A top 10 engine has being among the best . I always felt instinctly that a engine that is slightly weaker than Crafty must be within the top 10..
But now my test sugguests it's actually within the top 5 easily..

Based on what you write below, Yace is definitely among the "best", since it's almost equal to TCB.. right

I agree..What about Phalanx? I recall you mentioning on ccc that Phalanx is a very versatile program good at both Blitz and long time controls..Yace is even better at Blitz..



BTW what's new at the winboard forum? I stopped checking for a few days (busy organising my inter-university tournament) and i see a few new updates for
Francesa,Bringer, Techno and a Yace18h??..
Also what's with's the "pink" stuff?
Aaron

[Mogens Larsen]

Geschrieben von: / Posted by: Mogens Larsen at 25 June 2000 19:09:46:
Als Antwort auf: / As an answer to: Re: Yace against Phalanx. geschrieben von: / posted by: Aaron at 25 June 2000 18:29:49:

Okay different definition of "best". For you best doesn't mean top 10?
I see A top 10 engine has being among the best . I always felt instinctly that a engine that is slightly weaker than Crafty must be within the top 10..
Based on what you write below, Yace is definitely among the "best", since it's almost equal to TCB.. right
I agree..What about Phalanx? I recall you mentioning on ccc that Phalanx is a >very versatile program good at both Blitz and long time controls..Yace is even >better at Blitz..
Also what's with's the "pink" stuff?

Best means number one. Among the best means top five in my opinion.
Slightly weaker is your own opinion.
It would be a fair assumption that Yace is among the five or six best programs at blitz and therefore also among the ten best programs. Rather obvious.
I haven't tried Phalanx against Yace at longer timecontrols, but I would think that they're about even. But again, no evidence yet.
I've explained the purple stuff in another recent post to JH.
Best wishes...
Mogens

[Aaron]

Geschrieben von: / Posted by: Aaron at 25 June 2000 19:29:16:
Als Antwort auf: / As an answer to: Re: Yace against Phalanx. geschrieben von: / posted by: Mogens Larsen at 25 June 2000 19:09:46:

I always felt instinctly that a engine that is slightly weaker than Crafty >must be within the top 10..
Slightly weaker is your own opinion.

Based on what you write below, Yace is definitely among the "best", since >>it's almost equal to TCB.. right
I agree..What about Phalanx? I recall you mentioning on ccc that Phalanx is a >very versatile program good at both Blitz and long time controls..Yace is even >better at Blitz..
Also what's with's the "pink" stuff?

It would be a fair assumption that Yace is among the five or six best programs >at blitz and therefore also among the ten best programs. Rather obvious.
I haven't tried Phalanx against Yace at longer timecontrols, but I would think >that they're about even. But again, no evidence yet.
I've explained the purple stuff in another recent post to JH.

Yes..Of course..And at Blitz only..But i'm sure quite a few people share this opinion..I see about 2-3 lists that rank Yace as 2nd..

Thanks..So it's "among the best" ,meaning top 10 at least..
They are? I suspect Phalanx is stronger than Yace at longer time controls..But like you i have no evidence..Except what I see based on the battle of crowns calibration runs..

At Blitz..yace rules..

Sorry missed it..I'll look for it..

[Mogens Larsen]

Geschrieben von: / Posted by: Mogens Larsen at 25 June 2000 19:38:31:
Als Antwort auf: / As an answer to: Re: Yace against Phalanx. geschrieben von: / posted by: Aaron at 25 June 2000 19:29:16:

Yes..Of course..And at Blitz only..But i'm sure quite a few people share this opinion..I see about 2-3 lists that rank Yace as 2nd..
Thanks..So it's "among the best" ,meaning top 10 at least..
They are? I suspect Phalanx is stronger than Yace at longer time controls..But like you i have no evidence..Except what I see based on the battle of crowns calibration runs..

Yes, Yace is weaker that Crafty at blitz. Is Yace slightly weaker? Maybe.
What lists are you talking about?
Yes.
I think the strength difference is less significant than suggested by the BoC data, but Phalanx may still have an edge. I'll try running as many G/60 games as possible tonight.
Best wishes...
Mogens

[pete]

Geschrieben von: / Posted by: pete at 25 June 2000 20:02:18:
Als Antwort auf: / As an answer to: Re: Yace against Phalanx. geschrieben von: / posted by: Mogens Larsen at 25 June 2000 19:38:31:

I think the strength difference is less significant than suggested by the BoC data, but Phalanx may still have an edge. I'll try running as many G/60 games as possible tonight.
Best wishes...
Mogens

I think this might be one of the problems of the Calibration runs . Perhaps it is also a little about popularity of programs
regards.
pete

[Aaron]

Geschrieben von: / Posted by: Aaron at 25 June 2000 20:16:40:
Als Antwort auf: / As an answer to: Re: Yace against Phalanx. geschrieben von: / posted by: Mogens Larsen at 25 June 2000 19:38:31:

Yes, Yace is weaker that Crafty at blitz. Is Yace slightly weaker? Maybe.
What lists are you talking about?

Thanks..So it's "among the best" ,meaning top 10 at least..
They are? I suspect Phalanx is stronger than Yace at longer time controls..But like you i have no evidence..Except what I see based on the battle of crowns calibration runs..

Yes.
I think the strength difference is less significant than suggested by the BoC data, but Phalanx may still have an edge. I'll try running as many G/60 games as possible tonight.
Best wishes...
Mogens

Oh..You know those posts where people posts their "Blitz lists" rankings etc..I can hunt for the exact posts ,if you want , but it will take some time..
Wyx's list is one of them..of course..

[pete]

Geschrieben von: / Posted by: pete at 25 June 2000 20:17:15:
Als Antwort auf: / As an answer to: Re: Yace against Phalanx. geschrieben von: / posted by: pete at 25 June 2000 20:02:18:

I think the strength difference is less significant than suggested by the BoC data, but Phalanx may still have an edge. I'll try running as many G/60 games as possible tonight.
Best wishes...
Mogens

I think this might be one of the problems of the Calibration runs . Perhaps it is also a little about popularity of programs
regards.
pete

I should explain this a little better as else it sounds like an accuse which I don't intend to .
You think : yace is underrated.
I personally think Bringer is way underrated . Lot of Calibration games were played with Bringer1.5 for example .
So now ( if I would take part in the calibration process ) I could easily change the results to what i think is more true .
Hmm , how about 50 60/0 games against Phalanx ?
Like quite a few people know Phalanx always had problems with Bringer , so in my will to make the results a little "fairer" I flaw them .
This is the reason why I didn't send one single game of Bringer .
Best wishes .
pete

[Aaron]

Geschrieben von: / Posted by: Aaron at 25 June 2000 20:30:06:
Als Antwort auf: / As an answer to: Re: Yace against Phalanx. geschrieben von: / posted by: pete at 25 June 2000 20:17:15:

I should explain this a little better as else it sounds like an accuse which I >don't intend to .
You think : yace is underrated.
I personally think Bringer is way underrated . Lot of Calibration games were >played with Bringer1.5 for example .
So now ( if I would take part in the calibration process ) I could easily >change the results to what i think is more true .
Hmm , how about 50 60/0 games against Phalanx ?
Like quite a few people know Phalanx always had problems with Bringer , so in >my will to make the results a little "fairer" I flaw them .
This is the reason why I didn't send one single game of Bringer .

This sounds like something i hear when people discuss SDDF lists..people question Why some programs are matched only against weaker opposition etc etc

But this is impossible to avoid, unless, you regulate the same no of games against each opponent..
As it is..This is a mere calibration run..So as long as the engines play in approximately the correct "group" ..it's okay..Personally i feel Yace should be in group A, just to see how well it does at longer time controls..

[Mogens Larsen]

Geschrieben von: / Posted by: Mogens Larsen at 25 June 2000 20:34:14:
Als Antwort auf: / As an answer to: Re: Yace against Phalanx. geschrieben von: / posted by: Aaron at 25 June 2000 20:16:40:

Oh..You know those posts where people posts their "Blitz lists" rankings etc..I can hunt for the exact posts ,if you want , but it will take some time..
Wyx's list is one of them..of course..

No, that won't be necessary. There's been quite a few matchresults, but not that many rating lists based on a lot of games.
In Wyx's latest tournament Yace competed against the programs in group B, so the rating isn't based on results against the best engines. The next tournament will be more interesting when it comes to the capabilities of Yace.
Best wishes...
Mogens

[Aaron]

Geschrieben von: / Posted by: Aaron at 25 June 2000 20:38:13:
Als Antwort auf: / As an answer to: Re: Yace against Phalanx. geschrieben von: / posted by: Mogens Larsen at 25 June 2000 20:34:14:

Oh..You know those posts where people posts their "Blitz lists" rankings >>etc..I can hunt for the exact posts ,if you want , but it will take some >>time..
Wyx's list is one of them..of course..

No, that won't be necessary. There's been quite a few matchresults, but not >that many rating lists based on a lot of games.
In Wyx's latest tournament Yace competed against the programs in group B, so >the rating isn't based on results against the best engines.

Yes..There's two more I belive..But it's in the archives by now i belive..


Yes, i thought of the same thing..But even if 2nd place is a little too high, i don't think it will drop too far..

[Mogens Larsen]

Geschrieben von: / Posted by: Mogens Larsen at 25 June 2000 20:55:08:
Als Antwort auf: / As an answer to: Re: Yace against Phalanx. geschrieben von: / posted by: pete at 25 June 2000 20:17:15:

I should explain this a little better as else it sounds like an accuse which I don't intend to .
You think : yace is underrated.
I personally think Bringer is way underrated . Lot of Calibration games were played with Bringer1.5 for example .
So now ( if I would take part in the calibration process ) I could easily change the results to what i think is more true .
Hmm , how about 50 60/0 games against Phalanx ?
Like quite a few people know Phalanx always had problems with Bringer , so in my will to make the results a little "fairer" I flaw them .
This is the reason why I didn't send one single game of Bringer .
Best wishes .
pete

There's nothing to worry about Pete. Your initial statement sounded allright to me. Popularity, experience and excitement control the testing trend. Yace is an exiting engine at the moment. Last time around it was Amy. Who knows, the next big thing might be Bringer 1.7. I wouldn't mind that, because I like the professional look and feel about Bringer and the strength is increasing with every release.
You're also right about calibration. Some programs dislike playing against certain other programs and so on. The only solution is to play an exact number of games between every program, which is impossible in this case due to the nature of the contributions.
Best wishes...
Mogens

[Mogens Larsen]

Geschrieben von: / Posted by: Mogens Larsen at 25 June 2000 23:06:58:
Als Antwort auf: / As an answer to: Re: Yace against Phalanx. geschrieben von: / posted by: Aaron at 25 June 2000 20:38:13:

Yes..There's two more I belive..But it's in the archives by now i belive..
Yes, i thought of the same thing..But even if 2nd place is a little too high, i don't think it will drop too far..

Well, since you persist. I would like to know what rating lists you're referring too. Carlos Drake had one where Yace was 3rd, but the hardware may influence the results.
Make up your mind. Are you going to use fact or opinion? There's little point in starting with the intent of arguing by fact, just to change your mind when they aren't there. Yace played and won eight (8) games in group B, so you can't draw any conclusions.
Best wishes...
Mogens

[Dann [not glowing] Corbit]

Geschrieben von: / Posted by: Dann [not glowing] Corbit at 26 June 2000 02:59:59:
Als Antwort auf: / As an answer to: Re: Yace against Phalanx. geschrieben von: / posted by: Mogens Larsen at 25 June 2000 20:55:08:

I should explain this a little better as else it sounds like an accuse which I don't intend to .
You think : yace is underrated.
I personally think Bringer is way underrated . Lot of Calibration games were played with Bringer1.5 for example .
So now ( if I would take part in the calibration process ) I could easily change the results to what i think is more true .
Hmm , how about 50 60/0 games against Phalanx ?
Like quite a few people know Phalanx always had problems with Bringer , so in my will to make the results a little "fairer" I flaw them .
This is the reason why I didn't send one single game of Bringer .
Best wishes .
pete

There's nothing to worry about Pete. Your initial statement sounded allright to me. Popularity, experience and excitement control the testing trend. Yace is an exiting engine at the moment. Last time around it was Amy. Who knows, the next big thing might be Bringer 1.7. I wouldn't mind that, because I like the professional look and feel about Bringer and the strength is increasing with every release.
You're also right about calibration. Some programs dislike playing against certain other programs and so on. The only solution is to play an exact number of games between every program, which is impossible in this case due to the nature of the contributions.

Unfortunately, there are definitely going to be injustices with any procedures designed by humans. My "Battle of the Crowns" idea is designed to be self healing, however.
All the calibration games are used *only* to determine starting position. And if a program truly does get put in the wrong bracket, it will get promoted to the next bracket because it will dominate the lower one. Hence, for the next cycle, it will move to where it belongs. It is practically inconceivable that a two bracket error will be made except with the very bottom programs.


My FTP site