Any Operator for Stockfish @ ACCA

[swami]

I know it's late for announcement but better late than never. I forgot to make this post until I saw Eelco's call for Alaric operator. I see that Stockfish is eligible to play in the online events since Tord is the actively involved in the project himself.

I had initially asked Allard If he would be ok with Bright taking part in ACCA, he said Ok. I also offered to operate for Glaurung/Stockfish and contacted Tord the next day because I've a laptop, he said he wasn't particularly interested in any online tourneys but would not object if anybody wants to operate for his engine, and that he would gladly sign up the registration.

I wasn't aware of one operator - one program rule. So, I could no longer operate Stockfish in a core2Duo laptop. If someone has spare time in the weekend, they could always take it up. The engine is pretty strong and is actively being developed.

[Vinvin]

Great ! I really would like to see Stockfish playing this tournament !!

A Nehalem 4 or 8 cores would be great ! Any proposal ??

[CRoberson]

Sherif will operate it.

[CRoberson]

There hasn't been a one program per operator rule. The issue was when Allard would have two strong programs and you were going to operate one of
them and operate Stockfish (another very strong program). This situation would create many opportunities for teamwork which could manipulate
the tournament outcome.

So, I needed to adjust the rules to compensate for this problem which we have never had before.

Generally speaking, one operator can operate more than one program. However, at least one of the programs must be weak enough
so that this doesn't create potential for manipulating the tournament outcome.

The same rule will be in place for more than one program from the same author. However, I am going to allow Spark and Bright in
this tournament becuase Allard suggested a way to ensure there wouldn't be a problem. If Spark and Bright are paired against each other
(the bot handles this; not me or any other human), then the game will not be played and both will get 0 points for the round.

[swami]

Great! I had sent 2 e-mails to Sherif, and a pm just in case. He had finally took up the offer! I'm pleased.

[hgm]

However, I am going to allow Spark and Bright in
this tournament becuase Allard suggested a way to ensure there wouldn't be a problem. If Spark and Bright are paired against each other
(the bot handles this; not me or any other human), then the game will not be played and both will get 0 points for the round.

Sorry, but I think this is a ridicuous rule that will be much more disruptive for fair outcome of the tournament than the participation with two strong programs. So even if the rule does not apply to me, as a matter of principle I file a protest against it.

If you want to do anything along those lines, then at least both programs should get 1/2 point when they are paired. This will still be a disadvantage for the stronger program, and so presumably reduce the chances of winning of the tournament by any of the author's programs compaired to actually playing the game without cheating.

That it still might enhance the chances of the author to win compared to participating with only one program does not disturb me. This is a championship for engines, not for authors. And I think that is a good thing. There are more effective ways to cheat than writing a new second engine anyway.

[swami]

However, I am going to allow Spark and Bright in
this tournament becuase Allard suggested a way to ensure there wouldn't be a problem. If Spark and Bright are paired against each other
(the bot handles this; not me or any other human), then the game will not be played and both will get 0 points for the round.

Sorry, but I think this is a ridicuous rule that will be much more disruptive for fair outcome of the tournament than the participation with two strong programs. So even if the rule does not apply to me, as a matter of principle I file a protest against it.

If you want to do anything along those lines, then at least both programs should get 1/2 point when they are paired. This will still be a disadvantage for the stronger program, and so presumably reduce the chances

I'd not mind Bright getting 1/2 a point when it's paired against Spark. But 0 looks arbitrary.

It's 14 rounds tournament with less than 30 engines participation, So I'm pretty sure that Bright might face Spark at some point, despite Spark being 300 - 400 elo lower than Bright.

[Allard Siemelink]

Most probably, the Spark version that will participate is weaker than Bright, but not by as much as you presume.

The main concern was the opportunity for (accusations of) manipulation in case the engines have to play each other.
Not playing the game if two engines by the same author happen to get paired prevents that.

The other concern (by some) was 'unfair winning chances'.
Zero points for both would surely address this.
Of course, I would not mind 1/2 points each either and that too will reduce winning chances.

I think it is up to Charles to decide how high the penalty should be.

[hgm]

Assigning a default result by not playing the game is a matter that IMO is independent of assigning a penalty. Manipulation would be impossible whatever the result that would be assigned to the mutual game. I see no justification for an additional penalty.

One might as well assign penalties to engines that have too big a chance of winning beause their programmer has been working on it for such a long time already...

[Tord Romstad]

Sherif will operate it.

A few pedantic corrections to the information on the participants list:

• The engine name is "Stockfish", not "Stockfish-Glaurung". Perhaps "Stockfish-Glaurung" is just the ICC handle. This would be OK, of course.
• Joona Kiiski and Marco Costalba should be listed as authors (as well as myself, of course), not just as team members.
• It is not entirely true that Stockfish doesn't use any endgame tablebases or bitbases. It uses a single bitbase for KP vs K.