Opening Help

[Eelco de Groot]

Any chance of releasing Toga Mara? Should we create a begging thread?

Hi Ulysses,

I don't know if there is any chance. The last thing we heard from Taban was on October 1st in the thread I posted some details about, he had just found some new stuff in the search and was in a very good mood. In fact he talked about having created something like "a solution to mysterious chess".

And in response to my earlier post the answer was:

Eelco, if Vas dares put claim to any of my ideas/code,I'll dare him to decompile whichever Rybka n prove it. Anyway, he's a respectable guy who respects creativity.

But since then we have not heard a thing. I also said to Roy Brunjes who inquired, maybe contacting the author would teach us a little bit more. But I don't know if he reads on this forum and I believe all the Toga Mara threads on Rybka Forum have been relegated to the Flipside by Felix after the clones surfaced. Maybe a begging thread would help!

Or maybe we never hear from Toga Mara again...

Regards, Eelco

[Ovyron]

I should rephrase the question: What stops you from releasing Toga Mara Beta 6? Isn't it right if you just release it with the source? Is it worth it or is Toga Mara a weaker Toga/not improvement in style?

[Eelco de Groot]

Hi Ulysses,

Well I just don't think my version is really worth the effort yet of testing it. Apart from the fact that I'm using Taban's material.cpp and he has not released his Toga Mara yet.

I'm considering this version more as a 'shadow-version' of the Toga Mara. In case the story of Toga Mara turns out to be nothing more than that, that does not mean that Toga Mara could not be written. The mere fact that Rybka 3 is out there doing well, shows us it could be written If Taban does not do it we will just have to write it ourselves. Even if Toga Mara is from the future, the fact that we can even contemplate its existence makes it already real in a sense.

But to be honest creating a version to be as strong as Toga Mara is supposed to be would probably take me two dozen times 50 years or more to write, even with all the Toga contributions made by real programmers, like Michel van den Bergh who has done some work on the codebase recently.

The most valuable thing about this "Toga Mara Beta" version is probably that it has got multi-PV reestablished, with the help of Stanislav Tsukrov from ChessBase who fixed it when he made a port of Toga for the Pocket Fritz interface. It was lost in the Toga II 1.4 beta versions from Thomas Gaksch. The bugfix seems to work as far as I could see that is, and if there is any new version of Toga planned I think it can be included. (Bill Mowery may also have made a fixed multi-PV version already earlier than Stanislav but I have not checked his work myself)

Apart from Toga Mara which may or may not exist the most promising new version is I think Jerry Donald's new beta 12, which seems to have made some progress, at least some progress that is measurable, against Rybka 3 it is still a bit dismal I think. I hope it is okay to post some of the CCRL preliminary testsresults from Shaun:

Beta 12 is doing well in my gauntlet here but not enogh games yet, 239 of 600 games played, should have an update in a few days.

Toga II 1.4.3JDbeta12 1CPU - Bright-0.3d 1CPU 16.0 - 4.0 80.00%
Toga II 1.4.3JDbeta12 1CPU - Bright-0.4a 1CPU 12.5 - 7.5 62.50%
Toga II 1.4.3JDbeta12 1CPU - Deep Shredder 11 x64 1CPU 9.0 - 11.0 45.00%
Toga II 1.4.3JDbeta12 1CPU - Deep Sjeng 3.0 64-bit 1CPU 9.0 - 11.0 45.00%
Toga II 1.4.3JDbeta12 1CPU - Glaurung 2.1(JA) 64-bit 1CPU 11.5 - 8.5 57.50%
Toga II 1.4.3JDbeta12 1CPU - HIARCS 12 SP CCRL 10.5 - 9.5 52.50%
Toga II 1.4.3JDbeta12 1CPU - HIARCS Paderborn 2007 SP CCRL 11.0 - 9.0 55.00%
Toga II 1.4.3JDbeta12 1CPU - Loop M1-T 64-bit 1CPU 10.0 - 10.0 50.00%
Toga II 1.4.3JDbeta12 1CPU - Naum 3.1 64-bit 1CPU 9.0 - 11.0 45.00%
Toga II 1.4.3JDbeta12 1CPU - Rybka 3 64-bit 1CPU 2.5 - 17.5 12.50%
Toga II 1.4.3JDbeta12 1CPU - Spike 1.2 Turin 14.0 - 6.0 70.00%
Toga II 1.4.3JDbeta12 1CPU - Zappa Mexico II 64-bit 1CPU 8.0 - 11.0 42.11%
123.0 - 116.0 51.46%

You can check here for the beta11 results

Shaun

Also a certain way of increasing Toga's strength on big hardware would be the development of YBW as present in Cluster Toga. But at the moment that version is private, mainly because it is part of dissertation work from Kai Himstedt according to a post by Horst Wandersleben on CSS forum.
Clemens Keck posted a few details there, in German, about the work on Cluster Toga.

Eelco