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

BGonline.org Forums

More swiss tournaments, best in three shorter matches and clocks

Posted By: Maik Stiebler
Date: Wednesday, 15 October 2008, at 6:58 a.m.

In Response To: More swiss tournaments, best in three shorter matches and clocks (Bob Koca)

Player's skill changes over time and also the rating changes over time due to random fluctuations even if skill does not.

I agree with that and also with the bias argument you mention elsewhere. But I do not see how it matters.

Suppose we extract matches from the database where players were between 95 and 105 nominal elo points apart. We do not know a priori how large the actual average elo difference of the players is, but do we need to care? We can go ahead and construct METs from those matches and afterwards know what their actual average elo difference was.

Most of this thread was not about constructing METs, but about comparing the skill content of matches of different lengths. We could extract a set of M-point-matches and a set of N-point-matches, each with players ~100 nominal elo points apart. We would not know a priori how large the actual average elo difference of the players is, but couldn't we assume that, whatever it is, it is equal in both sets? I see there might be selection biases at work there, too, but that's not clear to me.

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.