2017-18 Stats/Rankings

[justinhub2003]

Originally Posted by sedziobs

Do they have a bracketology page? According to bracketmatrix's methodology, Dave Ommen at Bracketville is ranked #1. I don't see Screen the Screener or Mike Randle/ Gus Kearns at first glance though.

No they had a guest on that specific podcast. It was Jonathan warriner. I didn’t verify if it was true or not.


It’s just what they mentioned on the podcast.


But I do like this podcast

[ME80]

Andy Katz's Power 36 has UC in at 4 this week

[ervins84]

http://www.barttorvik.com/

Just stumbled on this because of Reddit. Figured anyone here would appreciate it.

Haven't checked it out completely, but the site has a lot of really good info and almost seems like a knock off kenoom that is completely free.

***UPDATE
Seriously can't believe this site is free. It's awesome.***

Best part is when you go under an individual team, you can tinker with projections and run simulations with however many fixed outcomes of games including conference tournament. But smartly, it's incorporating rpi tiers and rpi to come up with the seeds etc.

I could play around on this thing all day

[cincrulz11]

Originally Posted by ervins84

http://www.barttorvik.com/

Just stumbled on this because of Reddit. Figured anyone here would appreciate it.

Haven't checked it out completely, but the site has a lot of really good info and almost seems like a knock off kenoom that is completely free.

***UPDATE
Seriously can't believe this site is free. It's awesome.***

Best part is when you go under an individual team, you can tinker with projections and run simulations with however many fixed outcomes of games including conference tournament. But smartly, it's incorporating rpi tiers and rpi to come up with the seeds etc.

I could play around on this thing all day

Great find! This has some amazing stuff

[waterhead]

Originally Posted by ervins84

http://www.barttorvik.com/

Just stumbled on this because of Reddit. Figured anyone here would appreciate it.

Haven't checked it out completely, but the site has a lot of really good info and almost seems like a knock off kenoom that is completely free.

***UPDATE
Seriously can't believe this site is free. It's awesome.***

Best part is when you go under an individual team, you can tinker with projections and run simulations with however many fixed outcomes of games including conference tournament. But smartly, it's incorporating rpi tiers and rpi to come up with the seeds etc.

I could play around on this thing all day

Nice work! Very cool stuff here...

[sedziobs]

Originally Posted by ervins84

http://www.barttorvik.com/
seems like a knock off kenpom that is completely free.

Great find. Sort of merges Kenpom, rpiforecast, and Warren Nolan. I think each of those sites are better at what they do individually, but this one is a nice compilation. Thanks for sharing.

[sedziobs]

I've started to read into the methodology for Bart Torvik's T-Rank. He admits to copying Kenpom:

T-Rank is in the same league with Kenpom (no surprise, since I copied Kenpom liberally)

According to Massey's Rankings, T-Rank has the 4th best prediction success rate at 69.8%, just behind Kenpom's 69.9%. At #1 with 70.2% is a site called Singer, which is just a Google Docs spreadsheet that also uses Kenpom. This confirms in my mind that Kenpom is the best original algorithm. The other popular ones are Sagarin (ranked #30) and ESPN BPI (ranked #32).

The Teamsheet Ranks page lists the metrics at the top of the actual NCAA team sheet: http://warrennolan.com/basketball/20...eam=Cincinnati.
T-Ranketology simply lists the teams in order according to the "Overall" number at the top right of the team sheet. So, it's not really taking into account RPI Groups. It's still a nice unbiased automated tool.

[waterhead]

Originally Posted by sedziobs

I've started to read into the methodology for Bart Torvik's T-Rank. He admits to copying Kenpom:

T-Ranketology simply lists the teams in order according to the "Overall" number at the top right of the team sheet. So, it's not really taking into account RPI Groups. It's still a nice unbiased automated tool.

Interesting...last year our team sheet rank was #22 and we got a 6 seed. Those numbers equate very nicely. This year our team sheet rank (so far) is #5. That would be the #1 two seed.

Needless to say it will be interesting to see how the committee dismantles that resume...lol! Near as I can tell this is all current and not projected stuff.

[waterhead]

Originally Posted by waterhead

Interesting...last year our team sheet rank was #22 and we got a 6 seed. Those numbers equate very nicely. This year our team sheet rank (so far) is #5. That would be the #1 two seed.

Needless to say it will be interesting to see how the committee dismantles that resume...lol! Near as I can tell this is all current and not projected stuff.

The year before last our team sheet rank was #32 and we got a 9 seed. Again, this is equates almost perfectly. I guess you could quibble and say we deserved the last 8 seed.

[sedziobs]

Originally Posted by waterhead

Near as I can tell this is all current and not projected stuff.

There are actually three different T-Ranketologies. The one I described that's based on team sheets is current. He also has two others that combine a bunch of different factors, one of which is current and the other a projection. In those, he does include a system to handle RPI tiers:
1-25: 20 pts
26-50: 10 pts
51-100: 3 pts
101-200 loss: -1 pt
201+ loss: -4 pts

I'm not a big fan of this. I'm not sure why he wouldn't use the new RPI Groups. Plus, this system will tend to favor teams in better conferences.

I can't really complain though. There's a ton of good stuff on this site.