[Retros] A new program to solve fairy proof games

[CAILLAUD Michel]

Dear friends,
My feedback is the same as others expressed.
Super! Thanks you, François!
I met the same exception than Marco for positions where the "apparent"
number of played moves is much lower than the effective number of moves
played in the game.
I am not specialist but I wonder if the use of hash tables can help
(François mentioned hash keys, not the same thing??!). By experience I know
hash tables have spectacular influence on the solving time of these kinds
of position by Natch (orthodox). I don't know if Jacobi uses them (?!).
They are used in Euclide (fixed), Natch, Popeye and Gustav (dimensionable
by user).
Best,
Michel


<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
Garanti
sans virus. www.avast.com
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

On Wed, Nov 1, 2017 at 7:57 PM, Marco Bonavoglia <mbonavo at tin.it> wrote:

> On 30/10/2017 22:05, François Labelle wrote:
>
> Dear retro friends,
>
> I would like to introduce "Jacobi", a new program that can solve fairy
> proof games. The home page is at http://wismuth.com/jacobi/ . The program
> is a web app that runs in your browser and requires no installation.
>
> It can solve proof games that Popeye struggles with, which by itself is a
> good enough reason to add Jacobi to your toolbox. There are also new
> possibilities like allowing undefined pieces in the diagram of a proof
> game.
>
>
> Great program, thanks! I checked it with different fairy SPGs with
> excellent results, with one strange exception:
> PDB P1066705 (Initial position without BPh7 SPG4.0 Circe)
>
> Popeye solves it in abt 50 seconds, Jacobi takes a lot of time, after ten
> minutes it's at about 11% and has analyzed only 1. Sc3 1. Sa3 and was
> analyzing 1. Sh3 when I stopped it.
> I think it could take more than one hour. This on my laptop (with i7)
> using Chrome or Firefox (I tried both). Is it the homebase position that
> "confuses" Jacobi, or may be is a bug?
>
> Marco Bonavoglia
>
> _______________________________________________
> Retros mailing list
> Retros at janko.at
> https://pairlist1.pair.net/mailman/listinfo/retros
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://pairlist1.pair.net/pipermail/retros/attachments/20171101/df11cf2c/attachment.html>