Will TCEC need to rewrite their rules?

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

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
User avatar
xr_a_y
Posts: 789
Joined: Sat Nov 25, 2017 1:28 pm
Location: France

Re: Will TCEC need to rewrite their rules?

Post by xr_a_y » Mon Mar 11, 2019 6:56 am

gonzochess75 wrote:
Sun Mar 10, 2019 5:30 pm
Not interested in using this thread to defend Allie, but want to point out that any future rules you might come up with should account for the fact that Ethereal, Arasan, and Lc0 (probably a lot more?) are all using mostly the exact same (very important) code that was originally ported from Stockfish for tablebase support.
Nowadays, given the Fathom code, TB probing part in all engines can more or less be the same, and is very short. I nearly copy/paste Arasan one in Minic (and report a little issue in Jon's code at the same time). There is no "cloning" action in this, it is the community sharing ideas and tools. When a technology is good, everyone is using the same (if it is open ...). We could say the exact same thing for TT code, lazy SMP code, bitboard code, xboard or UCI protocol code, ...

jp
Posts: 841
Joined: Mon Apr 23, 2018 5:54 am

Re: Will TCEC need to rewrite their rules?

Post by jp » Mon Mar 11, 2019 5:44 pm

MikeB wrote:
Mon Mar 11, 2019 4:33 am
syzygy wrote:
Sun Mar 10, 2019 11:06 pm
gonzochess75 wrote:
Sun Mar 10, 2019 5:30 pm
Not interested in using this thread to defend Allie, but want to point out that any future rules you might come up with should account for the fact that Ethereal, Arasan, and Lc0 (probably a lot more?) are all using mostly the exact same (very important) code that was originally ported from Stockfish for tablebase support.
That code is not coming originally from Stockfish.
+1 ... how quickly people forget ...
If that's code just for TB support, the easy way out is to ban TB use for all engines.

Gian-Carlo Pascutto
Posts: 1191
Joined: Sat Dec 13, 2008 6:00 pm
Contact:

Re: Will TCEC need to rewrite their rules?

Post by Gian-Carlo Pascutto » Thu Sep 19, 2019 9:47 pm

Tablebases (and the associated access code) are considered an exception for very practical reasons. If using the third-party access code wasn't allowed, then every author who wanted to use EGTB would have to invent their own format (as Syzygy for example has no formal specification that you could program to) and the entities organizing the tournaments would somehow have to obtain the full set of tablebases in dozens of formats and a bag of HDD/SSD to store them on.

Despite this, barring bugs, the information contained and resulting play would be (largely) identical.

Unless you somehow manage to significantly improve over the Syzygy format, there is an extremely strong negative incentive to do all that. Practically, the result can be seen in TCEC16: the only engine who only supports an original EGTB format is running without EGTB because the TCEC machines don't have a copy of them. But it is true for end users as well: they're more likely to have a set of Syzygy.

Thus, trying to make an argument about clones/originality based on Syzygy access code can at best only be done in total ignorance of practical considerations, and at worst is just sheer intellectual dishonesty.

Disallowing all engines to use the Syzygy tables also works, but runs into the same practical objection: end users who care about strength are likely to at least have some of them, and nobody wants to watch engines play out an endgame that is solved. CCC makes the latter mistake and it's outright boring!

User avatar
Guenther
Posts: 3112
Joined: Wed Oct 01, 2008 4:33 am
Location: Regensburg, Germany
Full name: Guenther Simon
Contact:

Re: Will TCEC need to rewrite their rules?

Post by Guenther » Fri Sep 20, 2019 8:32 am

gonzochess75 wrote:
Sun Mar 10, 2019 5:30 pm
Not interested in using this thread to defend Allie, but want to point out that any future rules you might come up with should account for the fact that Ethereal, Arasan, and Lc0 (probably a lot more?) are all using mostly the exact same (very important) code that was originally ported from Stockfish for tablebase support.

...
Not sure if I regret it, but for once it needs to be spoken out:

1. Tablebase support (Syzygy - Scorpio - Gaviota - Nalimov too, if Eugene was contacted) was written for being used by all
2. It is not an important part (not even 15 rating points over a lot of games)
and it is easy for all TDs in the world to just let programs play w/o TBS

It seems pointing the finger at using the Syzygy implemenation of SF is nowadays just an excuse by some people to hide
some real 'stealing'... (not that I mean exactly you here Adam, but I am fed up reading this stupid argument)
Current foe list count : [101]
http://rwbc-chess.de/chronology.htm

Post Reply