| Subject: |
|
Re: Re: poor criticism |
| Name: |
|
Gino |
| Date Posted: |
|
Jun 4, 07 - 8:14 AM |
| Message: |
|
Hi Wim,
Nice graphs.
Regarding poor fit for tournaments with ratings < 2200 where players took less time, it is expected to see more variability given the increased percentual effect of the 24-hour buffer when time per move is smaller.
I can see benefit from linearizing your model (cumulative distribution curve) and comparing tournaments based on the slope of that curve, perhaps this trick could display not so obvious relationships. I wonder for example if the % of draws in an event has anything to do with time usage. I suspect it does.
Best,
Gino |
|
Replies:
|
|
|
|
|