Maybe you should try single-core next year

[Spock]

I agree with all of that, very sensible

[bob]

The most common argument of proponents for limited hardware was to not scare away amateurs from the tournament. It is not likely that more than two or three participants could afford clusters of 1000+ CPUs, and it is clear that a tournament with only two or three participants will not be viable.

For commercial programs, playing on large clusters is not very attractve, as their main customer base does not have such clusters. So it was coined that if a limit is to be imposed, at the current level of technology a quad would be a more logical choice than an octal. Quads are more or less standard now, in the consumer market.

It was discussed if it would not be better to require completely equal hardware, in stead of a limit on the number of cores. (And then have the organization provide the hardware, which would also solve the ever recurring problems tha accompany remote play.) Most participants felt that this was not strictly necessary, and a core limit would be enough. But it was not strongy rejected either, and is still considered an option.

As a drawback of the current hardware limit it was mentioned that the ability of using clusters is a major innovation of computer Chess, which would be suppressed by not awarding this effort in the WCCC. To circumvent this problem I suggested that we could limit the hadware to clusters of 4 nodes of 4 cores each. But this idea met with little approval. On the other side of this issue it was pointed out that allowing unlmited cluster size would detract effort from programmers to cluster programming, while there time would have been better spent by improving the single-CPU performance of their engine. Most ordinary users would not benefit from better scaling of a custer engine at all.

At some point it was mentioned that it was doubtful if Rybka would participate next year if the 8-core limit remained in force. That was enough to sway al other participants in favor of the limit!

I am not sure why you think a name change would be required, or what the extra M stands for. A cluster is not a computer, and a championhip for Humans usually does not allow the participation of teams, rather than individuals.

First, there is a bad assumption. The WCCC is _not_ an event for rank "amateurs". This is a world championship event, and there should be minimum standards for admission. There certainly have been such in the past where any unknown author had to submit games for analysis before they were accepted.

The WCCC is designed as a tournament among the _strongest_ computer chess players in the world. Not a subset of crippled programs running on limited hardware. The big iron is what made the early WCCC events interesting. If someone wants a limited hardware version, call it the old WMCCC or something, but keep the unlimited WCCC as it always has been.

If playing against a cluster scares someone off, they don't belong in the WCCC event anyway. They are always free to play in other events including the online ones.

The point about a cluster/team is irrelevant. A cluster is a single computer from the perspective of a cluster-aware chess engine, it is not a "team" of different players. The "M" was for "Microcomputer" and was the microcomputer event for many years, limited with respect to hardware that could be used. The WCCC was never limited until the insane decision this year. They _thought_ it would increase participation. They fail to understand that the computer hardware issue is _not_ what is killing the WCCC. It is the cost/time required to participate that is driving it into the cellar.

[bob]

Well, I think two separate tournaments had the potential to be a good compromise for everyone. But as I said in my earlier post, the hardware limited section still had hugely expensive hardware (the W5580s) as well as a huge hardware differential. The exact same concerns that were supposed to have been addressed by the hardware limit but clearly weren't. So it wasn't thought through properly. I still think two separate ones has the potential to keep everyone happy, but the hardware limited one should be equal hardware (perhaps someone like Dell or HP may be prepared to loan some identical quad boxes) or a lower affordable hardware limit such as quads. Of course as technology progresses it would need to be reviewed. And the overly aggressive competitive attitudes from some commercial team representatives needs to be toned down, in my opinion.

uniform platform has been tried and found wanting. What processor do you use? X86? Or X86-64? Or PPC? or Alpha? Or ARM? Or something custom like the belle/deep thought days? Any limit only reduces the quality of the opponents, which is exactly the opposite of what you want for a World Championship event.

I'd prefer a uniform book as opposed to a uniform hardware platform. I'm not going to invest all of my time in book cooks as the commercial guys do.

[bob]

Based on the discussions, in my opinion, I think that they should retain the current setup where there is a WCCC and an Olympiad so as to satisfy both sides.

Maybe WCCC would have different categories like:

Uniform Hardware
No Hardware Limit

Uniform Hardware would make sure that other teams would not have an unfair advantage of having much faster and expensive machine just like what happened in the recent WCCC where other teams used inferior hardware. The hardware to be used must be the most common hardware used by ordinary people.

No Hardware Limit would showcase the combination of both hardware and software innovations.

If I can afford the expenses I will try to join WCCC next year with my amateur engine.

Before you write such a post, you need to give some _serious_ thought to what you are suggesting.

For example, a commercial entity can buy a big cluster, and use it 24/7 to prepare book cooks for the real tournament. No way you can limit that. Which means money can still buy a tournament just as surely as exceptional hardware can produce a win when used to play the actual games. But one can also use exceptional hardware to _prepare_ for the real games even if it is not used _in_ the games.

There is _no_ way to have a reasonable competition where money does not help.

To think otherwise is not thinking at all.

And if you discount the importance of the book cooking that goes on, again, you have not been involved enough to know how important the book actually is.

[bob]

A limit of 12-plies, maybe Rybka won't win? I'll be interesting to see what the anti-Rybka camp comes up with next.

That would be a winner of an idea. When you can't even trust the depth/nps/etc that Rybka reports.

[bob]

A limit of 12-plies, maybe Rybka won't win? I'll be interesting to see what the anti-Rybka camp comes up with next.

Funny, 12 Rybka plies mean 14-15 Fritz plies and 15-17 Junior plies. Rybka will win with 100% score. The idea is ridiculous, as programmers will start tweaking ply depth showed.

Kai

"start"???

[Nid Hogge]

The most common argument of proponents for limited hardware was to not scare away amateurs from the tournament. It is not likely that more than two or three participants could afford clusters of 1000+ CPUs, and it is clear that a tournament with only two or three participants will not be viable.

For commercial programs, playing on large clusters is not very attractve, as their main customer base does not have such clusters. So it was coined that if a limit is to be imposed, at the current level of technology a quad would be a more logical choice than an octal. Quads are more or less standard now, in the consumer market.

It was discussed if it would not be better to require completely equal hardware, in stead of a limit on the number of cores. (And then have the organization provide the hardware, which would also solve the ever recurring problems tha accompany remote play.) Most participants felt that this was not strictly necessary, and a core limit would be enough. But it was not strongy rejected either, and is still considered an option.
.

Oh come on. You don't grasp the point. This is the WCCC. Just like Bob said. It's supposed to have the strongest players possible. It's the "Champions league" of Computer Chess. But hey, FC Barcelona and Manchester United are so much better than these "amateurs" teams out there, maybe they should start the game with 8 players? They also got bigger budget.

This analogy can be made so many times it's not even funny.
Perhaps Usain Bolt should run with only one leg, Michael Phelps could swim blindfolded, and Michael Shumacher will drive with one wheel on his Ferrari F1.

Right?

Equal hardware is just as worse. The whole point of "Sportsmanship" is that you go out and try to compete with some other guys that might be better than you. What's the point of forcing everyone to limit they're max potential? Some are better. That's the whole beauty of an 'underdog'. Rybka is better because it's smarter. It has more knowledge of some sort, because someone added that knowledge. A 2000 ELO engine with 30,000 cores will still finish last if it's dumb. That's not the case with Rybka, Shredder, Sjeng and the rest out there. these people started from nowhere in order to achieve something. And after all the work they dedicate someone comes up and puts a big FU in their face. Why?

The future is multicore, and if the ICGA doesn't like it, well, honsetly, sorry to say this, it's their friggin' problem. It's not like it was not anticipated or came in a surprise. the world+dog knew it was coming. So they should have made that senseless decision already back in 2004, not in 2009 where multicore systems are already widespread. But the ICGA didn't know a thing, and has just landed from outer space.. welcome aboard.

Your'e stifling innovation and common nature in the name of competition.
There is no excuse for crippling. Not in CC, BB, Soccer, or any other sport. If there was one, you would be living in an absurd world where better individuals are punished because they are better. Way to go.

[Laskos]

Your'e stifling innovation and common nature in the name of competition.
There is no excuse for crippling. Not in CC, BB, Soccer, or any other sport. If there was one, you would be living in an absurd world where better individuals are punished because they are better. Way to go.

That is all fine, but what if one comes with Blue Gene/Q supercomputer? Would it be fair?

Kai

[Leto]

Your'e stifling innovation and common nature in the name of competition.
There is no excuse for crippling. Not in CC, BB, Soccer, or any other sport. If there was one, you would be living in an absurd world where better individuals are punished because they are better. Way to go.

That is all fine, but what if one comes with Blue Gene/Q supercomputer? Would it be fair?

Kai

Absolutely it would be fair, and if they win with that supercomputer I'd consider them the Computer Chess Champion. To me the title of Computer Chess Champion means strongest chess computer entity, not the strongest software, we already know what the best software is thanks to CEGT and CCRL etc.

[Edsel Apostol]

Based on the discussions, in my opinion, I think that they should retain the current setup where there is a WCCC and an Olympiad so as to satisfy both sides.

Maybe WCCC would have different categories like:

Uniform Hardware
No Hardware Limit

Uniform Hardware would make sure that other teams would not have an unfair advantage of having much faster and expensive machine just like what happened in the recent WCCC where other teams used inferior hardware. The hardware to be used must be the most common hardware used by ordinary people.

No Hardware Limit would showcase the combination of both hardware and software innovations.

If I can afford the expenses I will try to join WCCC next year with my amateur engine.

Before you write such a post, you need to give some _serious_ thought to what you are suggesting.

For example, a commercial entity can buy a big cluster, and use it 24/7 to prepare book cooks for the real tournament. No way you can limit that. Which means money can still buy a tournament just as surely as exceptional hardware can produce a win when used to play the actual games. But one can also use exceptional hardware to _prepare_ for the real games even if it is not used _in_ the games.

There is _no_ way to have a reasonable competition where money does not help.

To think otherwise is not thinking at all.

And if you discount the importance of the book cooking that goes on, again, you have not been involved enough to know how important the book actually is.

What I'm basically suggesting is to keep both format so as to satisfy both sides. In my opinion, if one's engine is stronger it will dominate whatever the hardware is. It doesn't matter if it's uniform single core or mega cluster. That's the point I'm trying to prove. Take a look at the result in Pamplona, Rybka dominates whatever event, limited or unlimited hardware.

About the book cooking you've mentioned, I'm aware that it's an essential part of the tournament. I already knew that way before when I was just reading computer chess programming theory and doesn't have an engine yet. I didn't mention in my post that I discount the importance of a book. My beta tester/operator/book maker is actually cooking some book for my engine already.

Yes, I have not been involved enough but that doesn't necessarily mean that I knew nothing.