Stockfish @ WCCC - good or bad idea?

1215 views
Skip to first unread message

timoha...@gmail.com

unread,
May 1, 2015, 5:15:23 AM5/1/15
to fishc...@googlegroups.com
Dear all,

some time ago there was a thread in talkchess forum about a possible Stockfish participation in the WCCC. Some members of the Stockfish community (i.e. Marco, Tord, Lucas, ...) expressed that they aren't interested in this event at all, since it is of no statistical relevance. They clearly stated that an official registration of the Stockfish team in ICGA events will never happen. But the bottom line was that there would be no problem whatsoever if somebody else would play with Stockfish in the WCCC, as long as it is beyond question that the Stockfish team has nothing to do with it. At least that was my interpretation of the discussion.

Day before yesterday I got a reminder e-mail from ICGA that the registration deadline for the ICGA events is on May 1st, 2015. After that entry fee would be doubled. Since I'm lucky to have vacation at the end of June / beginning of July I thought it would be fun to travel to Leiden and watch the ICGA events. But there was one thing I was missing since Rybka disappeared from ICGA events: The world's strongest program didn't show up there any more. It felt like a human world championship without Magnus Carlsen. So I asked the ICGA people if somebody already registered Stockfish and in case nobody did so, if they would accept my registration. I clearly stated that it is no participation request from "official Stockfish team", but only from somebody who would call himself a computerchess enthusiast and would like to see at least one of the world's two strongest chess programs participate in the tournament. I told them that there was a discussion about that a few weeks ago in talkchess forum and that somebody stated that GPL allows that and gives explicit permission. Their first reply was very positive and they told me that they would like to see me operating Stockfish in the tournament. Some things still have to be clarified (e.g. entry fee), but if that can be resolved it seems that they would accept my registration.

But of course I don't want to play there against the will of the Stockfish community. If somebody has strong objections against this I would of course withdraw the registration. Please keep in mind that the ICGA-WCCC is just one tournament among many others in modern times. 15-20 years ago it had great importance in the (computer) chess world, but these times are over. For me it would be great fun to be part of it regardless. I've been travelling to over-the-board tournaments for over 10 years now, mainly IPCCC (visitor since 1999, Spike operator in 2007), WCCC (2006 Spike, 2008 ClusterToga), CSVN tournaments (2010 Deep Sjeng, 2014 Hannibal) and Thuringen (2004-2011, different programs). I really enjoy seeing high class chess games on a real chess board, talking to the opponent's operator and watching the player's "thoughts" on the screens. It's a positive mix between tension (of course I'm engrossed by the match and root for the program I'm operating) and relaxation (just to sit at the board and pursue my hobby). I also like watching (computer) chess online (e.g. TCEC), but for me it is greater fun to be on site and watch the games on a real chess board. Since I thought it would be a good idea to finally see the strongest program play in world championship, I had the idea of doing something by myself to see that happen.

But of course I don't want to annoy somebody who contributed much more to the Stockfish project than me (I only shared 20347 CPU hours [User: team-oh], that's a ridiculous small part compared to the thousands of hours the developers [mainly Tord, Marco and Joona] spend in front of their screens writing program code). So please tell me your thoughts about it.

Best regards
Timo

jo.mu...@gmail.com

unread,
May 1, 2015, 6:05:11 AM5/1/15
to fishc...@googlegroups.com
Bad idea.

(1) SF taking the title from the likes of Junior and Shredder is absolutely meaningless. The only true touchstone for SF is K (who cannot be bothered to show up).

(2) SF taking part in an ICGA event should grant to much of an honour to people who handled the Rybka case in such a biased and dictatorial manner, disregarding the fundamental achievement that computer chess, SF and K including, owes to Rajlich.

By the way, I have only recently started to donate CPU time.

Kamyar

unread,
May 1, 2015, 6:40:35 AM5/1/15
to fishc...@googlegroups.com
There is also the issue of hardware. You didn't talk about that. Surely ICGA is a terrible idea, but even if it wasn't, the operator would need to bring the pc that runs the games as well. The directors aren't so nice to provide high end hardware like in TCEC.

You'll probably have more fun if you attend the ACG (advances in computer games) seminar in Leiden university instead. Here is the link https://acg2015.wordpress.com

lilmafya

unread,
May 1, 2015, 7:03:33 AM5/1/15
to fishc...@googlegroups.com, timoha...@gmail.com
I personally would like to see SF competing in all the tournaments,
so I would vote for: "Go for it"!! 
:)

lucas....@gmail.com

unread,
May 1, 2015, 7:20:33 AM5/1/15
to fishc...@googlegroups.com, timoha...@gmail.com
The driving principle of open source is: do it yourself. So, arguing or voting about it won't make it happen. Someone will need to go go to Leiden for a week, pay an entry feed to the ICGA, bring their own hardware, and participate in this tournament. As you can already guess, that someone certainly won't be me...

timoha...@gmail.com

unread,
May 1, 2015, 8:56:52 AM5/1/15
to fishc...@googlegroups.com, jo.mu...@gmail.com
I can understand your arguments, but still...

1) Nobody knows if Komodo will show up there. A few weeks ago I heard from a reliable source that they were searching for a possibilty to rent a computer there. Probably they were also searching for an operator. This must be no proof that they really will participate, but it could be a clue.

2) The Rybka case is a sad story and everyone has his own opinion about it. But from my point of view this has nothing to do with Stockfish.

timoha...@gmail.com

unread,
May 1, 2015, 9:00:00 AM5/1/15
to fishc...@googlegroups.com
Am Freitag, 1. Mai 2015 12:40:35 UTC+2 schrieb Kamyar:
> There is also the issue of hardware. You didn't talk about that. Surely ICGA is a terrible idea, but even if it wasn't, the operator would need to bring the pc that runs the games as well. The directors aren't so nice to provide high end hardware like in TCEC.
>
> You'll probably have more fun if you attend the ACG (advances in computer games) seminar in Leiden university instead. Here is the link https://acg2015.wordpress.com

I know that I have to care for the hardware by myself. That should be no problem since for the main tournament (unlimited hardware) the computer must not be on site - it can also be accessed remotely.
Message has been deleted

lucas....@gmail.com

unread,
May 1, 2015, 9:02:25 AM5/1/15
to fishc...@googlegroups.com, timoha...@gmail.com, jo.mu...@gmail.com
No need to convince me. My opinion is that the ICGA is as useless as it is arrogant, and that calling their tournament a "world" championship is total usurpation.

But, some people think differently, and would like SF to participate. If one of those people is willing to participate, I don't see why we should prevent him.

timoha...@gmail.com

unread,
May 1, 2015, 9:04:00 AM5/1/15
to fishc...@googlegroups.com, timoha...@gmail.com, lucas....@gmail.com
Am Freitag, 1. Mai 2015 13:20:33 UTC+2 schrieb lucas....@gmail.com:
> The driving principle of open source is: do it yourself. So, arguing or voting about it won't make it happen. Someone will need to go go to Leiden for a week, pay an entry feed to the ICGA, bring their own hardware, and participate in this tournament. As you can already guess, that someone certainly won't be me...

You are absolutely right: Someone has to do it to make it happen - that was what I thought and took action. Of course it's clear that I have to take care for travelling and accomodation costs for myself as well as the entry fee and the hardware. I'm aware of this and that's no problem for me at all.

timoha...@gmail.com

unread,
May 1, 2015, 9:05:06 AM5/1/15
to fishc...@googlegroups.com, timoha...@gmail.com
Am Freitag, 1. Mai 2015 13:03:33 UTC+2 schrieb lilmafya:
> I personally would like to see SF competing in all the tournaments,
> so I would vote for: "Go for it"!! 
> :)
>
Thanks, but for the moment I only plan to participate in the main tournament (with unlimited hardware). I don't like Blitz, so the speed tournament is definitely out. The WCSC could still be discussed.

cwsto...@gmail.com

unread,
May 1, 2015, 10:24:37 AM5/1/15
to fishc...@googlegroups.com, timoha...@gmail.com
While i also only contributed CPU time, personally i'm also all for SF competing at the WCCC - provided it has competitive hardware. Except for Komodo, no engine comes close to Stockfish these days so everyone will expect it to win the WCCC in champion style. Do you already have an idea what kind of setup you'd be going there with (in case you do)? Oh and btw, thanks for showing initiative :)

Lux Aeterna

unread,
May 1, 2015, 10:55:23 AM5/1/15
to fishc...@googlegroups.com
On Friday, 1 May 2015 12:40:35 UTC+2, Kamyar wrote:
the operator would need to bring the pc that runs the games as well.

Not having bothered delving too much into the subject before, this detail surprises me. Participants bring their own hardware? As in, each engine is allowed to run on its own hardware, possibly perfectly tailored to suit that engine? Like that, I'd even be willing to enter SF into the championship; running a severely tweaked Linux on a 16-core box for pure performance. I saw some dramatic performance improvements when going from Windows to Linux and then to a low latency kernel.

Seriously, if every participant in the WCCC is allowed to run on its own hardware, the tournament is even more of a joke than it appears to be.

cwsto...@gmail.com

unread,
May 1, 2015, 11:42:44 AM5/1/15
to fishc...@googlegroups.com
Am Freitag, 1. Mai 2015 16:55:23 UTC+2 schrieb Lux Aeterna:
> Like that, I'd even be willing to enter SF into the championship; running a severely tweaked Linux on a 16-core box for pure performance. I saw some dramatic performance improvements when going from Windows to Linux and then to a low latency kernel.
>
> Seriously, if every participant in the WCCC is allowed to run on its own hardware, the tournament is even more of a joke than it appears to be.

Ok, but you might have to compete with the likes of this:

"At the WCCC 2011 Jonny ran on a a cluster of 100 PCs with eight cores each, and at the WCCC 2013 in Yokohama, even with 2400 cores of the new btrzx3 cluster of the University of Bayreuth , where Jonny was quite close to become champion - if it had won the last round versus Junior."

Not sure how strong Jonny is on a multiple-thousand core cluster compared to SF on 16 cores, but for example on IPON, there's a 383 ELO gap between the strongest versions of Jonny and SF.

Then there's also the WCSC where (according to wiki) competing programs must run on machines with identical hardware specifications which in 2013 was an Intel quad core i7, 2.7 GHz, 16MB Hash. 16MB seems kinda small to me tbh, but whatever. It would be nice for SF to win those titles, but TCEC is more important imo.
Message has been deleted
Message has been deleted

Lux Aeterna

unread,
May 1, 2015, 12:36:04 PM5/1/15
to fishc...@googlegroups.com, timoha...@gmail.com
Well, given the fact some really shitty engines have been winning the WCCC, engines that are absolutely demolished by SF, in a level playing field, I am confident SF would not need that many cores to play on their level.

Irrelevant to that though is my actual point -- each participant being allowed its own hardware invalidates the result and the validity of the WCCC. Basically, it would amount to having human players setting their own time limit and increment on the clock in a tournament. No one would accept such conditions, so no one should accept any such conditions in a software chess tournament.

SNP_Ti...@hotmail.com

unread,
May 1, 2015, 12:52:20 PM5/1/15
to fishc...@googlegroups.com, timoha...@gmail.com
On Saturday, May 2, 2015 at 2:36:04 AM UTC+10, Lux Aeterna wrote:
> Well, given the fact some really shitty engines have been winning the WCCC, engines that are absolutely demolished by SF, in a level playing field, I am confident SF would not need that many cores to play on their level.
>
>
> Irrelevant to that though is my actual point -- each participant being allowed its own hardware invalidates the result and the validity of the WCCC. Basically, it would amount to having human players setting their own time limit and increment on the clock in a tournament. No one would accept such conditions, so no one should accept any such conditions in a software chess tournament.

The WCCC isn't a software chess tournament, a fact which those that run it are well aware of, which is why there is a parallel competition, the World Chess Software Championship, which requires all engines to run on the same hardware.

pasquale....@gmail.com

unread,
May 1, 2015, 1:00:39 PM5/1/15
to fishc...@googlegroups.com, timoha...@gmail.com
@timo: IMO is always a good idea when a polite guy wants to have fun with his hobby. I will cheer for you (and for SF, obviously) at WCCC

On Friday, 1 May 2015 11:15:23 UTC+2, tea...@gmx.de wrote:
> For me it would be great fun to be part of it regardless. (...) I really enjoy seeing high class chess games on a real chess board, talking to the opponent's operator and watching the player's "thoughts" on the screens. (...) but for me it is greater fun to be on site and watch the games on a real chess board.
> Best regards
> Timo

timoha...@gmail.com

unread,
May 1, 2015, 2:23:35 PM5/1/15
to fishc...@googlegroups.com, pasquale....@gmail.com, timoha...@gmail.com
Am Freitag, 1. Mai 2015 19:00:39 UTC+2 schrieb pasquale....@gmail.com:
> @timo: IMO is always a good idea when a polite guy wants to have fun with his hobby. I will cheer for you (and for SF, obviously) at WCCC
>

Thank you for that friendly statement! I hope that SF will play some nice games there. Especially I'm looking forward to the game between Jonny and Stockfish: Jonny on a big cluster with 1000 or more cores, Stockfish on just one computer (with 8, 16 or max 32 cores). I saw Jonny play in Nov 2014 in Leiden and this was quite an impressive performance there (7 points out of 7). Jonny really crushed all opponents there, but this time it will get better competition...

Lux Aeterna

unread,
May 1, 2015, 2:33:58 PM5/1/15
to fishc...@googlegroups.com, timoha...@gmail.com, SNP_Ti...@hotmail.com
On Friday, 1 May 2015 18:52:20 UTC+2, SNP_Ti...@hotmail.com wrote:
The WCCC isn't a software chess tournament

Very true; in the format described, it is not even a tournament. It's a tech demo. Well, I prefer my tech demos more graphical in nature. ;)

Marco Costalba

unread,
May 1, 2015, 7:01:09 PM5/1/15
to timoha...@gmail.com, fishc...@googlegroups.com
Personally I am against it. 

I know igca required people operating engines to be authors or strictly related to them. Now it seems they have relaxed this rule and introduce a special case. 

Well, this tells a lot about their hypocrisy... 

vdb...@gmail.com

unread,
May 1, 2015, 9:01:07 PM5/1/15
to fishc...@googlegroups.com, timoha...@gmail.com
On Friday, May 1, 2015 at 7:01:09 PM UTC-4, Marco Costalba wrote:
> Personally I am against it. 
>
>
> I know igca required people operating engines to be authors or strictly related to them. Now it seems they have relaxed this rule and introduce a special case. 
>
>
> Well, this tells a lot about their hypocrisy... 
>

No it doesn't. Stockfish is a special case since it is GPL. Anyone can be an author of SF. That's a situation they never encountered before.

lucas....@gmail.com

unread,
May 1, 2015, 10:16:32 PM5/1/15
to fishc...@googlegroups.com, vdb...@gmail.com, timoha...@gmail.com
Sure it would be nice to see SF kurb-stomp all these patzer programs that pretend to be world champion material.

But the format of this tournament has some serious issues:
* different hardware, different opening books => it's a hardware/book war more than fair engine war!
* statistical relevance: it's an "all play all" format. Yes, that means engine A plays engine B, but you don't even get a rematch B vs. A with colors reversed. And with 5-6 engines participating, you can imagine the little stastical relevance of the result. May as well toss a coin to decide the winner.

And there are some technicalities:
* someone has to go to Leiden for a week, bring his hardware, pay an entry fee to the ICGA, and spend a week with these people.
* that someone needs the approbation of all the people who contributed to SF and therefore have equal copyright prerogatives. Personally, I don't care and will not oppose my vote, but it only takes one to block the process.

I know people want to see SF compete, but the ICGA does not deserve this honor. We should ignore them and let them turn into a fossil. Oh wait! They're already a fossil, maybe they'll turn into petrol now.

TCEC is coming soon, and that's a far more interesting and relevant. Anyone with a half-brain understands that it is the real world championship, and WCCC is only a marketing event to save legacy commercial engines from falling into oblivion.

vdb...@gmail.com

unread,
May 1, 2015, 10:48:21 PM5/1/15
to fishc...@googlegroups.com, vdb...@gmail.com, lucas....@gmail.com, timoha...@gmail.com

> * that someone needs the approbation of all the people who contributed to SF and therefore have equal copyright prerogatives. Personally, I don't care and will not oppose my vote, but it only takes one to block the process.


This is complete and utter nonsense. The only one that could keep SF from competing in an ICGA tournament is the ICGA itself. There is nothing in the GPL license that restricts the use of software.




lucas....@gmail.com

unread,
May 2, 2015, 2:59:26 AM5/2/15
to fishc...@googlegroups.com
@vdb: you're saying that anyone can enter with SF, without the permission or knowledge of its authors. Of course the ICGA creates the rules they want, but allowing that would be ridiculous. It would show how desperate they are (trying to compensate the lack of participants).

timoha...@gmail.com

unread,
May 2, 2015, 3:38:51 AM5/2/15
to fishc...@googlegroups.com, timoha...@gmail.com
Am Samstag, 2. Mai 2015 01:01:09 UTC+2 schrieb Marco Costalba:
> Personally I am against it. 
>
>
> I know igca required people operating engines to be authors or strictly related to them. Now it seems they have relaxed this rule and introduce a special case. 
>
>
> Well, this tells a lot about their hypocrisy... 
>

Hi Marco,

thank you for your statement. Seems like I misunderstood the ICGA-thread (http://talkchess.com/forum/viewtopic.php?t=55395) on talkchess back then. My impression was you don't care if SF plays there because you aren't interested in the ICGA. But ok, now you made it clear. Personally I'm disappointed about your decision because now the situation will stay the same as before (world championship without "Magnus Carlsen"). But of course I will be true to my own word: I won't participate there with Stockfish against the will of the authors.

So I will cancel my registration of Stockfish on Monday.

Good luck @ TCEC,
Timo

Marco Costalba

unread,
May 2, 2015, 4:11:23 AM5/2/15
to timoha...@gmail.com, fishc...@googlegroups.com
I have always written all the worst that is possible to write about igca and the farce they call world  championship, you can find my comments with broad explanations everywhere, just Google for it. 

Now I stay consistent with that : I don't want to endorse them, I don't want they use  SF to advertise themselves and their void tournament. 

vdb...@gmail.com

unread,
May 2, 2015, 4:39:17 AM5/2/15
to fishc...@googlegroups.com, lucas....@gmail.com
On Saturday, May 2, 2015 at 2:59:26 AM UTC-4, lucas....@gmail.com wrote:
> @vdb: you're saying that anyone can enter with SF, without the permission or knowledge of its authors. Of course the ICGA creates the rules they want, but allowing that would be ridiculous. It would show how desperate they are (trying to compensate the lack of participants).

You are turning things upside down. It would be completely ridiculous for the ICGA to require that each of the hundreds of SF-copyright holders gives individual permission for SF to enter.

In fact by putting their contributions under the GPL the copyright holders have _already_ given legal permission for SF to be used in any way which is compatible with the GPL. And there is nothing in the GPL that would prohibit SF from entering the ICGA tournament.


pasquale....@gmail.com

unread,
May 2, 2015, 4:53:33 AM5/2/15
to fishc...@googlegroups.com, timoha...@gmail.com
@timo: your action speaks for yourself. Chapeau.
Message has been deleted

Lux Aeterna

unread,
May 2, 2015, 5:34:20 AM5/2/15
to fishc...@googlegroups.com, timoha...@gmail.com
On Saturday, 2 May 2015 09:38:51 UTC+2, timoha...@gmail.com wrote:
now the situation will stay the same as before (world championship without "Magnus Carlsen").

The thing you have to remember though is that the ICGA is neither officially sanctioned nor recognized by anyone as the governing body concerning machine chess (be it software or hardware). In other words, their championship is not an actual world championship. It is just called that, nothing more.

vdb...@gmail.com

unread,
May 2, 2015, 5:57:22 AM5/2/15
to fishc...@googlegroups.com, timoha...@gmail.com

>
> The thing you have to remember though is that the ICGA is neither officially sanctioned nor recognized by anyone as the governing body concerning machine chess (be it software or hardware).

The ICGA _is_ affiliated with the FIDE.

https://www.fide.com/fide/directory/affiliated-organizations.html?comid=44&task=committee

So they do have official recognition.

Lux Aeterna

unread,
May 2, 2015, 6:40:34 AM5/2/15
to fishc...@googlegroups.com, vdb...@gmail.com, timoha...@gmail.com
It does not quite work that way; being recognized by a recognized entity does not imply recognition. More important memberships and affiliations to look at would be the IOC, the ARISF and so forth. Seriously, ICGA is NOT the defining entity as far as qualifying a tournament as the official world computer chess championship is concerned.

lilmafya

unread,
May 2, 2015, 8:42:18 AM5/2/15
to fishc...@googlegroups.com, timoha...@gmail.com, vdb...@gmail.com
So if you decided to participate, then the 3 factors to win is :
1) Engine Strength
2) Hardware used
3) Opening book

In case you decided to go for it you can contact me at my e-mail and I can provide you with the opening book Fauzi 4.1.abk (Which is private), since I really would like SF to win it, and for that we need to look at all the aspects.
Fauzi.abk book is the strongest .abk book out, made for tournaments and correspondence games, choosing always the strongest line and many time leading to victory without even going out of book.

Thanks and regards,

pasquale....@gmail.com

unread,
May 2, 2015, 8:58:06 AM5/2/15
to fishc...@googlegroups.com, timoha...@gmail.com, vdb...@gmail.com
To make a short story long:

About WCCC: talking about a world championship we are talking about an organization who organize a championship between his members scattered in the world (i.e. between people who recognize some value to the organization): FIFA World Cup, FIA F1 World Championship etc.
The moniker "WCCC" is a shortcut to "ICGA WCCC" and it has the same *questionable marketing value* of the moniker "World Series" for the "MLB World Series"


Regarding the polite Timo's request: Lucas and Marco don't want recognize any value to ICGA, and IMO the personal feelings of the SF's authors are more valuable for the SF Project than a formal discussion about GPL and/or ICGA rules or a victory in a tournament.


Regarding SF and heterogeneous tournament: I have nothing against a computer chess tournament with custom hardware, custom software, custom opening book. If the chess algorithms could be massively parallelized on GPU, SF, in the actual form, will be a thing of the past.


Regarding questionable tournament rules: in order to improve SF, any single SF's loss is more valuable than thousand victories. With SF in the TOP 2 and a massive gap with others engines, there are very few losses to analyze (Komodo 9 and SF matches seem to be very drawish).


Regarding chess games and heterogeneous tournament: I have fun watching interesting chess games, and sometimes interesting chess games happens between imbalanced opponents within questionable rules (as Kasparov-Short in St. Louis the past week).

Ralph Stoesser

unread,
May 3, 2015, 1:26:40 AM5/3/15
to fishc...@googlegroups.com
If I would have the time I would do it and take all the costs, but only in case of an objection from one of the original authors. Reason is simple. If all of them would agree with the participation I wouldn't want to take all the costs myself. Except maybe if they would change the id String to something that reflects the multi-authorship better.
Reply all
Reply to author
Forward
0 new messages