Where did fargo ratings get all their data from? I am listed and have a rating but I don't remember ever playing in an event that uses fargo ratings or submits results to fargo rating.
If your rating is 400, 525 or 625 then you've been assigned the "starter" rating for former Leisure, Open and Advanced division players respectively.
If it's something different from one of the starter ratings, I'm fairly certain they're using your league results and adjusting your rating from the results of your matches against your opponents that have [more] established Fargo ratings and, to a lesser degree, the results of your opponents against
other opponents with [more] established Fargo ratings.
My concern, which I voiced to Mark Griffin in a FB conversation, is that the extrapolation of data is going to get REALLY thin in leagues where there are few, or even no, players with established Fargo ratings.
In my league, for example, the most established Fargo rating has a "robustness" of only 113. In fact, only 3 players have a rating "robustness" of over 20. The robustness of my rating, for instance, is only 15.
As I understand it (mind you, I'm fairly familiar with ELO ratings of this sort), EVERYONE ELSE'S ratings are going to be extrapolated from their results in their league matches against the 3 "established" players mentioned above (along with their results against me and the few other low-robustness rated players) and, to a lesser degree, from their results against other players who have played the 3 "established" players.
There are data smoothing techniques that have proven effective for working with relatively small sample sizes, such as my league's results, but the possibility of an outlier increases exponentially as the data set grows smaller.
My experience has shown me that the biggest negative effect that outliers have is that they lead (and are already leading in this case) to a general distrust of the formula by pool players who aren't familiar with the law of big numbers; which is to say 99% of them.
That being said, I'm hugely optimistic about the use of Fargo ratings at the BCAPL Nationals and in pool at large. I just hope that enough of the players can overcome their initial wariness with a math-based rating system, and overlook the initial glitches in what will prove to be a very small percentage of the ratings, long enough to let Fargo become more established and therefore, more accurate.
Taek Chang
)
