Arimaa Forum - Print Page


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

        Arimaa >> General Discussion >> Arimaa rating on BGG
        
(Message started by: omar on Jan 14^th, 2011, 12:15pm)

Title: Arimaa rating on BGG
Post by omar on Jan 14^th, 2011, 12:15pm

I thought I had some idea of how the 'Geek Rating' for games are calculated on BGG. Essentially about 100 5.5 votes are added to the average user rating to get the 'Geek Rating'. So I was surprised to see that Arimaa having 193 votes with an average user rating of 7.63 was getting a Geed rating of 6.805 while Bughouse Chess having 190 votes with an average user rating of 7.49 was getting a Geek rating of 6.807.

http://boardgamegeek.com/abstracts/browse/boardgame

Here is the explanation I had found:

Quote:

If you look at a list of games within BGG you may see a BGG Rating number between 1 and 10. If you then click on that game and go to the game's main page you will see a User Rating and an Average Rating. Most of the time, the BGG Rating number will be different from the User and Average Ratings. Why is this?
The User Rating of a game is the Average Rating rounded down to 1 decimal point. The Average Rating for a game is the average of all ratings from registered BGG users that the game has received, calculated by adding up all individual ratings and dividing by the number of ratings.
The BGG Rating is based on the Average Rating, but the number is altered. BoardGameGeek's ranking charts are ordered using the BGG Rating. To prevent games with relatively few votes climbing to the top of the BGG Ranks, artificial "dummy" votes are added to the User Ratings. These votes are currently thought to be 100 votes equal to the mid range of the voting scale: 5.5. The effect of adding these dummy votes is to pull BGG Ratings toward the mid range. Games with a large number of votes see their BGG Rating alter very little from their Average Rating, but games with relatively few user ratings will see their BGG Rating move considerably toward 5.5. This is known as "Bayesian averaging" and a quick search of both BGG and/or the Web will reveal much discussion on the topic.
In effect, usually the games with many votes will Rank higher than those games with the same Average Rating but fewer votes.

but the Geek rating difference between Arimaa and Bughouse Chess don't seem to fit this.

Does anyone know more about how the Geek ratings are calculated?

Title: Re: Arimaa rating on BGG
Post by Adanac on Jan 14^th, 2011, 12:51pm

on 01/14/11 at 12:15:44, omar wrote:

Omar, the values that you're quoting are for the "Abstract Games" sub-section which uses a different formula than the Overall rating. The geek rating for the two games is:

Bughouse Chess 6.278
Arimaa 6.201

I've heard they're also planning to add (in fact, they've probably already done so) a time-decay factor to the weightings now so that older votes don't count as much as new votes. They show the date that you voted for the game and you can refresh it at any time.

By the way, that "Abstract Game Ranking" list is a joke. Chess is ranked # 49 behind Blokus (ranked # 12 !!??), Mahjong, Santorini, Skat, Cribbage, TAMSK, Billabong, etc. It's probably good that Arimaa is flying below the radar and mostly hard-core fans are voting.

Title: Re: Arimaa rating on BGG
Post by omar on Jan 15^th, 2011, 11:37pm

Thanks Greg. I didn't know there was another Overall rating. I guess they like to keep the formula kind of mysterious. Anyway it represents only the view of the BGG community. If the 2 million+ members of chess.com decided to vote, I'm sure they could push chess to #1 :-)

Title: Re: Arimaa rating on BGG
Post by megamau on Jan 16^th, 2011, 12:44am

The effect you are seeing is that of "anti-shilling" algorithm.

If you think about it, with a normal bayesan average it is impossible that Arimaa could be lower than Bughouse. In fact it has:

* An higher "pure" average rating
* An higher number of votes (thus moderating the effect of the 5.5 "dummy" ratings)

The anti-shilling mechanism tries to prevent people to "cheat" the system by fake ratings both in positive and in the negative. The design is not pubblicized for the very same purpose (otherwise people would cheat the ant-shill).

However a reasonable guess it is that it modifies the very extreme ratings (1 and 10) which may have been given for shilling and it favours a more normal Gauss distribution.

Of course lifestyle games like arimaa, chess, checkers are hit by this mechanism....usually if you like them you LOVE them and they are bound to have many 10s.