|
Post by cyberVBmidwest on Mar 27, 2007 18:05:04 GMT -5
As the only combined division ranking system for college level men's volleyball, I was wondering what fans thought of the rankings and their predictions on who is most likely to win a match? I have not performed an analytical review of its success but I do recall seeing results that seemed to hold true of the rankings. There are a couple of matches that might slightly skew the rankings such as BYU & PSU versus Cal Baptist. However, we have a significant number of matches including cross divisional matches to smooth out the math. Please post your thoughts? MHPR Combined Division Rankings
|
|
|
Post by The Bofa on the Sofa on Mar 27, 2007 20:44:27 GMT -5
Do my thoughts count? ![:)](//storage.proboards.com/forum/images/smiley/smiley.png) Here's what I can tell you: Of approximately 850 matches that I have in the database, 238 of them are interdivisional. The combined Pablo rankings linked above correctly predict 209 of those 238 matches, which is about 87.8%. Overall (about 860 matches), the Pablo ratings currently get about 89.3% right, so Pablo is doing a little better on intradivisional matches (89.8% for intradivisional). The interdivisional matches mentioned above include DI vs DII. If I take out the D1 vs D2, it drops to 87.5%. However, that also includes the BYU and Penn St vs Cal Baptist matches. If I throw out those matches, the interdivisional success rate is 88.5%. These are all really close, so there is no indication that Pablo is doing significantly worse for interdivisional matches than it does within a division. Note, however, that I am only talking about the analysis in retrospect. I have not tracked the results of PREdiction, although I don't expect them to be all that different. Overall, Pablo has been getting a successful prediction rate of about 85% on average so far. In fact, Pablo has been more accurate than it usually is for women's DI.
|
|
|
Post by swingaway on Mar 28, 2007 12:13:26 GMT -5
I like it. I think 85% is unreal and shows a very well written formula. What a great effort by middle hitter and p-dub.
I need to come up with an easyer way next year (if there is a next year)to have Pablo compete in the PTW as a baseline for everyone.
|
|
|
Post by VBbeast on Mar 29, 2007 9:31:42 GMT -5
In my mind anything above 80% is a pretty good system particularly given the relatively small amount of data and the many nuances to an open division sport like men's volleyball.
Thumbs up!
|
|
|
Post by BeachBoy on Mar 29, 2007 21:57:05 GMT -5
Has BS me seen these rankings?
|
|