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

BGonline.org Forums

Maybe it's time for a change?

Posted By: Rick Janowski
Date: Friday, 19 February 2016, at 9:42 p.m.

In Response To: Maybe it's time for a change? (Florin Popa)

MWC cost is useful for evaluating which of the two players played better in a match but it does not naturally yield an absolute parameter like PR where all players may be considered together. Moreover, the results from a single match for PR or MWC cost are subject to significant variation and are therefore reliable.

It would be great if the principles of MWC cost could be used to give an alternative PR measure, but I think we need Xavier to do that for us with perhaps with the help of someone dealing with the mathematics.

However, in practice, rarely is a PR difference outcome significantly different than MWC Cost in terms of differential equity.

VR Elo as a measure will tend to be less reliable to VR long-term winrate, because recent results have exponentially greater influence than older results. (this is assuming the win rate is normalised to deal with different match lengths and average opponent strengths).

A study of online league results (all play all) with large sample size carried out within last 12 months did not demonstrate that long term win rate had better correlation with VR winrate than with PR. The opposite appeared to be the case.

It has been assumed that VR adjusted win rate is more reliable than PR, but never demonstrated in practice. Maybe the way XG measures luck is too crude for this approach.

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.