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

BGonline.org Forums

Ken's proposal

Posted By: Rod
Date: Wednesday, 18 December 2013, at 2:33 p.m.

In Response To: Ken's proposal (Matt Cohn-Geier)

I agree to a point. PR will be an accurate predictor. But I believe it doesn't tell the whole story. We've argued it before. I won't bother again. But at its heart, I believe that XG could achieve better results against human opponents if it could take opponent strength into account. We would then consider its opponent-strength-based plays to be perfect 0.0 PR.

Yes, someone with a 4 avg PR will, over time, outperform someone with a 5. But a 3.8 won't necessarily outperform a 4.1.

Not only that but one can, instead of being lucky with a win streak, be lucky (or unlucky) with PR in the short term. But that dissipates over time.

I tend to "band" players. By that, I mean, group them, at the low end, into 0.5 PR groupings and 1 PR groupings and 2 PR groupings when the play gets worse than 10. The slightest luck even over a very long time has such an outsized effect that a 0.2 PR difference is virtually meaningless. And I band them for purposes of play consideration.

Messages In This Thread

 

Post Response
Your Name:
Your E-Mail Address:
Subject:
Message:

If necessary, enter your password below:

Password:
 

 

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

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