[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums

Ken's proposal

Posted By: Matt Cohn-Geier
Date: Wednesday, 18 December 2013, at 9:27 a.m.

In Response To: Ken's proposal (Rod)

There tends to be pretty strong correlation between expected win rate and past results. Looking over the table, it is likely that the top 3 players: Dana, Neil, and Mochy, are better than the bottom 3: Herb Roman, Bill Riles, and Ben Elliott. I agree that using past performance as a metric is better than throwing random darts blindly.

However, there are big outliers, and there is almost no way to differentiate between players who are close. Take spots #6-10 (me, Alfred Mamlet, Rod, Stick, and Ed O'Laughlin) and it is not at all obvious that they are in the right order, or are actually worse than the top 5 or whatever. Still likely to be better than the bottom 10 players, but then again, there might be a random Falafel or Petko strewn in there.

Overall, the correlation between playing well and achieving results tends to be strong. But when you are looking at data aggregated from hundreds of coin flips, sometimes there will be streaks and outliers.

The biggest problem with PR is not that is not an accurate predictor. I think PR is a very accurate predictor of results. Sure, one can construct scenarios where PR is less than completely accurate, but it clearly is a better predictor than going off of past results. The problem is that there isn't that much data, the data that there is is hard to find, even if you do it might be cherry picked and have bias, etc.

Messages In This Thread


Post Response

Your Name:
Your E-Mail Address:

If necessary, enter your password below:




[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums is maintained by Stick with WebBBS 5.12.