Arimaa Forum - Print Page


    
      
        Arimaa Forum
        (http://arimaa.com/arimaa/forum/cgi/YaBB.cgi)
      

        Arimaa >> Arimaa Buzz >> Pour La Science Magazine
        
(Message started by: omar on Jul 1^st, 2010, 4:54am)

Title: Pour La Science Magazine
Post by omar on Jul 1^st, 2010, 4:54am

A French user mentioned to me that the rating system used on the Arimaa site was discussed in this magazine which has an article this month about rating systems.

http://www.pourlascience.fr/ewb_pages/c/ce-mois-ci.php

Title: Re: Pour La Science Magazine
Post by Fritzlein on Jul 1^st, 2010, 6:18am

Apparently one must subscribe in order to see the whole article. I am curious whether the Arimaa rating system they talk about is the one used in the game room, or whether Arimaa is merely mentioned as a site that uses Coulom's Whole History Ratings.

In any event, the mention seems to have gleaned us a couple of new French registrations.

Title: Re: Pour La Science Magazine
Post by Arimabuff on Jul 1^st, 2010, 10:53am

What's funny is that "Pour la science" is actually the French version of "Scientific American". So chances are that you already have it in your libraries.

Title: Re: Pour La Science Magazine
Post by Janzert on Jul 1^st, 2010, 3:58pm

Looks like this is the cover article for Pour la Science but the cover article for Scientific American this month is "The Universe is Leaking" and doing a scan of the contents I don't see anything about rating systems. So I doubt it's in Scientific American.

Janzert

Title: Re: Pour La Science Magazine
Post by woh on Jul 3^rd, 2010, 7:31am

Arimaa is only mentioned for using WHR.
To his knowledge it is the only place where his WHR system is used.

Title: Re: Pour La Science Magazine
Post by Fritzlein on Jul 3^rd, 2010, 9:38am

Great, thanks for the update, woh. I suppose, then, that Coulom will be very interested in our experience with WHR, because it is his only test case. :)

Title: Re: Pour La Science Magazine
Post by Fritzlein on Jul 9^th, 2010, 3:48am

Oops, I totally missed that Coulom was the author of the article rather than one of its subjects. Thanks for enlightening me, leo. So does he himself discern any flaws in WHR, or does he present it as the pinnacle of ratings achievement and the state of the art?

Title: Re: Pour La Science Magazine
Post by leo on Jul 10^th, 2010, 12:07pm

The paper is a history of rating systems, and Rémi Coulom presents his Whole History Rating as one of the three systems with the best predictive power together with TrueSkill Through Time and EDO. But he readily admits that a single number will never be enough to describe somebody's skill :) If I remember from his online paper, we're still a far cry from 100% predictive power for any system.

Pour la science and Scientific American share articles but PLS isn't exactly the French version of SA in spite of the subtitle, which explains why the contents often differ greatly.