TCEC stage 3 , New Houdini starts with a bang

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Harvey Williamson, bob

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
mjlef
Posts: 1421
Joined: Thu Mar 30, 2006 12:08 pm
Contact:

Re: TCEC stage 3 , New Houdini starts with a bang

Post by mjlef » Sun Jul 24, 2016 2:21 am

Nay Lin Tun wrote:
shrapnel wrote:Also, the complaint was made 4 to 5 hours into the game. Complaint should have been dismissed. It should have been made earlier.
I'm sure nothing would have been said if Komodo was actually winning !
The Conspiracy theorists just got a big boost !
Yes of course, complaints always come from the losers not from the winners.If the organisers would like to blame the wrong settings, that settings must be in public before the game start! Who knows K team would like to try with the new setting that was used in the previous game and only when the game was dead lost, the drama started. A couple of things come to my mind
1. Complaints should be done before the game start and should be invalid when the game has decisive outcome
2. No transparency in private setting
3. The setting were not obviously wrong as a general sense{eg,people can sense as a wrong setting if only 1 core was used instead of 20}
4.No one knows whether the other engine authors might complaints due to the wrong setting or other reason and whether those complaints were refused.
5. The replay rule was already announced before the game but it was supposed to be an obvious bug , and Komodo did not play as a bug game.(there was intense fight in the whole game until one side got advantage)

P.S-->there was a public poll by someone whether the game should be replayed or not, and most spectators voted as no.
http://www.strawpoll.me/10816599/r
How can complaints be made before the game starts? The settings are not visible until the game starts.

I did not know Komodo was playing in the first game. I was at a Movie at 3 pm European Time when the game started. After the game, I ate, came back to the hotel, and called Larry a few times, did some compiles. I did not check TCEC until all that was done.

When I did check TCEC, I saw Komodo's score was for white a bit over 1. I checked the settings by right clicking the gear and immediately so no Contempt = 7 setting, and that Dynamism was set to 110 instead of the default 115 for Komodo 10.1

I emailed Anton, Martin and Larry to tell them about the wrong settings. Anton answered as follows (my email under Anton's reply):

>>
Dear Mark,

I am CC-ing Martin, this must be replayed if what you say was not set.

2016-07-23 21:11 GMT+03:00 Mark L. <xxxxx>:
Martin and Anton,

I looked at the TCEC settings for Komodo and they look wrong. We asked that Contempt be set to 7, and it is not mentioned (so I assume it defaults to 10). And the Komodo 10.1 setting for Dynamism is 115, not 110.

Can you please fix them?

Mark
>>

I never asked for the game to be replayed. Anton knew immediately is should be due to the settings errors.

The Komodo team had nothing to do with the wrong settings. We did not switch any programs

Taking a poll serves no purpose. TCEC has rule and plays fair. I respect them for that.

mjlef
Posts: 1421
Joined: Thu Mar 30, 2006 12:08 pm
Contact:

Re: TCEC stage 3 , New Houdini starts with a bang

Post by mjlef » Sun Jul 24, 2016 2:27 am

IanO wrote:
carldaman wrote:What was wrong with Komodo's settings? People should at least be entitled to a proper explanation by the organizers.
The differences can be found in the initial game comment.

Before:

Code: Select all

BlackEngineOptions&#58; Use Syzygy=true; Syzygy Probe Depth=2; Syzygy Probe Limit=6; Dynamism=110;
After:

Code: Select all

BlackEngineOptions&#58; Contempt=7;
To summarize, removing custom Syzygy probe options and "Dynamism" (whatever that is), and adding Contempt 7.

It would be nice to hear from Mark what the effects of these changes are likely to be.
Contempt = 7 is pretty important. We lowered the Contempt for this stage because the average opponent rating is higher than the last stage. Using a higher Contempt would have encouraged Komodo to try to avoid draws harder, and to keep more pieces on the board at some sacrifice of positional play. Dynamism also was wrong and our testing showed in Komodo 10.1, the higher setting was better.

TCEC has been very accurate at settings in the past, but we are all only human and some errors can occur. When they are found it should be corrected and replayed, which is what they did, and swiftly too.

There will be lots of games in this Stage.

gotogo
Posts: 91
Joined: Tue Jan 26, 2010 4:03 am

Re: TCEC stage 3 , New Houdini starts with a bang

Post by gotogo » Sun Jul 24, 2016 2:50 am

TCEC nepotism for you. TCEC had more than ample time to set all engines up correctly!

I'm sure Robert won't be happy, furthermore you guys may have just ruined it for everyone with Houdini!

I hope for a Stockfish and Houdini final!
Everyone was excited then the game was took from Houdini!

APassionForCriminalJustic
Posts: 415
Joined: Sat May 24, 2014 7:16 am

Re: TCEC stage 3 , New Houdini starts with a bang

Post by APassionForCriminalJustic » Sun Jul 24, 2016 4:58 am

gotogo wrote:Komodo team steals from Houdini!
I guess The TCEC sponsors are happy about outcome!

Who knows if TCEC changed some setting in Houdini?
Does Komodo retain game info in hash and change strategy after a loss?
Could this have been a setup all along?
As much as ARB is a pain in the @ss he now has some merit!

Why is it Mark always gets his way when he whines?
Does TCEC find games that favor Komodo over other engines?
I don't trust TCEC outcome anymore?
The operator didn't have the settings correct. There is no conspiracy... some of the settings affect Komodo ELO-wise by quite a bit... the operator is to blame for NOT ensuring that Komodo’s settings were correct. As far as I'm concerned... the first game was a fluke anyways... why couldn't houdini repeat another victory? Haha.

APassionForCriminalJustic
Posts: 415
Joined: Sat May 24, 2014 7:16 am

Re: TCEC stage 3 , New Houdini starts with a bang

Post by APassionForCriminalJustic » Sun Jul 24, 2016 5:00 am

gotogo wrote:TCEC nepotism for you. TCEC had more than ample time to set all engines up correctly!

I'm sure Robert won't be happy, furthermore you guys may have just ruined it for everyone with Houdini!

I hope for a Stockfish and Houdini final!
Everyone was excited then the game was took from Houdini!
Time or not, tcec screwed up... and what is this houdini to the final crap? Houdart only just started improving houdini... we all know that Komodo is stronger... get some brains...

APassionForCriminalJustic
Posts: 415
Joined: Sat May 24, 2014 7:16 am

Re: TCEC stage 3 , New Houdini starts with a bang

Post by APassionForCriminalJustic » Sun Jul 24, 2016 5:02 am

mjlef wrote:
IanO wrote:
carldaman wrote:What was wrong with Komodo's settings? People should at least be entitled to a proper explanation by the organizers.
The differences can be found in the initial game comment.

Before:

Code: Select all

BlackEngineOptions&#58; Use Syzygy=true; Syzygy Probe Depth=2; Syzygy Probe Limit=6; Dynamism=110;
After:

Code: Select all

BlackEngineOptions&#58; Contempt=7;
To summarize, removing custom Syzygy probe options and "Dynamism" (whatever that is), and adding Contempt 7.

It would be nice to hear from Mark what the effects of these changes are likely to be.
Contempt = 7 is pretty important. We lowered the Contempt for this stage because the average opponent rating is higher than the last stage. Using a higher Contempt would have encouraged Komodo to try to avoid draws harder, and to keep more pieces on the board at some sacrifice of positional play. Dynamism also was wrong and our testing showed in Komodo 10.1, the higher setting was better.

TCEC has been very accurate at settings in the past, but we are all only human and some errors can occur. When they are found it should be corrected and replayed, which is what they did, and swiftly too.

There will be lots of games in this Stage.
People are just desperate for a houdini miracle... Houdini will fail like it did before with houdini 4... while Houdart was on vacation both Komodo and Stockfish walked away with the crown.

User avatar
Laskos
Posts: 9409
Joined: Wed Jul 26, 2006 8:21 pm
Full name: Kai Laskos

Re: TCEC stage 3 , New Houdini starts with a bang

Post by Laskos » Sun Jul 24, 2016 5:05 am

IanO wrote:
carldaman wrote:What was wrong with Komodo's settings? People should at least be entitled to a proper explanation by the organizers.
The differences can be found in the initial game comment.

Before:

Code: Select all

BlackEngineOptions&#58; Use Syzygy=true; Syzygy Probe Depth=2; Syzygy Probe Limit=6; Dynamism=110;
After:

Code: Select all

BlackEngineOptions&#58; Contempt=7;
To summarize, removing custom Syzygy probe options and "Dynamism" (whatever that is), and adding Contempt 7.

It would be nice to hear from Mark what the effects of these changes are likely to be.
The values "Before" were close to to optimal. The ELO loss was no larger than 5-10 ELO points and hardly the reason for the lost game. Seems unfair to Houdini to replay the game.

carldaman
Posts: 1707
Joined: Sat Jun 02, 2012 12:13 am

Re: TCEC stage 3 , New Houdini starts with a bang

Post by carldaman » Sun Jul 24, 2016 5:13 am

Laskos wrote:
IanO wrote:
carldaman wrote:What was wrong with Komodo's settings? People should at least be entitled to a proper explanation by the organizers.
The differences can be found in the initial game comment.

Before:

Code: Select all

BlackEngineOptions&#58; Use Syzygy=true; Syzygy Probe Depth=2; Syzygy Probe Limit=6; Dynamism=110;
After:

Code: Select all

BlackEngineOptions&#58; Contempt=7;
To summarize, removing custom Syzygy probe options and "Dynamism" (whatever that is), and adding Contempt 7.

It would be nice to hear from Mark what the effects of these changes are likely to be.
The values "Before" were close to to optimal. The ELO loss was no larger than 5-10 ELO points and hardly the reason for the lost game. Seems unfair to Houdini to replay the game.
That was my reasoning as well.

User avatar
Laskos
Posts: 9409
Joined: Wed Jul 26, 2006 8:21 pm
Full name: Kai Laskos

Re: TCEC stage 3 , New Houdini starts with a bang

Post by Laskos » Sun Jul 24, 2016 5:33 am

carldaman wrote:
Laskos wrote:
IanO wrote:
carldaman wrote:What was wrong with Komodo's settings? People should at least be entitled to a proper explanation by the organizers.
The differences can be found in the initial game comment.

Before:

Code: Select all

BlackEngineOptions&#58; Use Syzygy=true; Syzygy Probe Depth=2; Syzygy Probe Limit=6; Dynamism=110;
After:

Code: Select all

BlackEngineOptions&#58; Contempt=7;
To summarize, removing custom Syzygy probe options and "Dynamism" (whatever that is), and adding Contempt 7.

It would be nice to hear from Mark what the effects of these changes are likely to be.
The values "Before" were close to to optimal. The ELO loss was no larger than 5-10 ELO points and hardly the reason for the lost game. Seems unfair to Houdini to replay the game.
That was my reasoning as well.
To see how disproportionate the action to replay the game is, for at most 5-10 ELO points loss in ONE game, they changed the result in the whole Stage 3 by 1 point between Komodo and Houdini, and that in 56 games each amounts to 13 ELO points in the WHOLE stage.

Uri Blass
Posts: 8554
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: TCEC stage 3 , New Houdini starts with a bang

Post by Uri Blass » Sun Jul 24, 2016 7:46 am

Laskos wrote:
IanO wrote:
carldaman wrote:What was wrong with Komodo's settings? People should at least be entitled to a proper explanation by the organizers.
The differences can be found in the initial game comment.

Before:

Code: Select all

BlackEngineOptions&#58; Use Syzygy=true; Syzygy Probe Depth=2; Syzygy Probe Limit=6; Dynamism=110;
After:

Code: Select all

BlackEngineOptions&#58; Contempt=7;
To summarize, removing custom Syzygy probe options and "Dynamism" (whatever that is), and adding Contempt 7.

It would be nice to hear from Mark what the effects of these changes are likely to be.
The values "Before" were close to to optimal. The ELO loss was no larger than 5-10 ELO points and hardly the reason for the lost game. Seems unfair to Houdini to replay the game.
I do not think that we know the elo loss at long time control and there is no basis to claim that the wrong setting is not the reason for the lost game.

I see nothing unfair for Houdini.

Post Reply