Around the Country

[justinhub2003]

Originally Posted by waterhead

I agree it would be nice to get a constant stream of rankings of RPI and Kenpom for teams either in or straddling lines of column 1 or 2...but that takes some effort. There is probably an easy way to do this but I am not a spread sheet genius...lol!

Those teams outside of column 1 or 2 aren't going to mean all that much to us in the end.

In terms of RPI (my least favorite metric) Jerry Palm used to have a tool called the Palm reader and when you clicked a specific team, it showed you in the side bar ways your team could improve its RPI that day by games on the schedule (teams you beat and teams you lost to)

[waterhead]

Originally Posted by justinhub2003

In terms of RPI (my least favorite metric) Jerry Palm used to have a tool called the Palm reader and when you clicked a specific team, it showed you in the side bar ways your team could improve its RPI that day by games on the schedule (teams you beat and teams you lost to)

I will continue to say this...RPI doesn't mean squat until the end when it starts to match the other metrics. But in the end it means a lot to the committee but they don't really evaluate this stuff until the end anyway.

Up until the last month or so Kenpom is better to look at while RPI is getting narrowed down.

[waterhead]

Originally Posted by waterhead

I will continue to say this...RPI doesn't mean squat until the end when it starts to match the other metrics. But in the end it means a lot to the committee but they don't really evaluate this stuff until the end anyway.

Up until the last month or so Kenpom is better to look at while RPI is getting narrowed down.

I don't recall our final RPI and Kenpom last year leading into the tourney but they were a lot closer than #7 and #34. They were probably closer to dead on around #12 or so within a few spots...again IDK.

[sedziobs]

Warren Nolan's site has automated team sheets, separating the schedule into the four RPI groups. The Bearcats' team sheet is here: http://warrennolan.com/basketball/20...eam=Cincinnati

There is also a "Predicted Team Sheet" that gives us a total of five Group 1 wins at the end of the season. However, the predictions are deterministic, meaning the most likely result of each individual game is used. The deterministic approach has UC going 28-3 with the only loss at Wichita. Kenpom uses probabilistic predictions, which give us a 26-5 record.

Just to clarify further for non-nerds: say we play 10 games with a 60% chance of winning each. Our deterministic predicted record is 10-0. The probabilistic prediction is 6-4. Determinism gives poor forecasts, so it is likely that the RPIs in Warren Nolan's predicted team sheet will look quite different. It's still an interesting way to get a feel for what our resume may show in March.

[waterhead]

Originally Posted by sedziobs

Warren Nolan's site has automated team sheets, separating the schedule into the four RPI groups. The Bearcats' team sheet is here: http://warrennolan.com/basketball/20...eam=Cincinnati

There is also a "Predicted Team Sheet" that gives us a total of five Group 1 wins at the end of the season. However, the predictions are deterministic, meaning the most likely result of each individual game is used. The deterministic approach has UC going 28-3 with the only loss at Wichita. Kenpom uses probabilistic predictions, which give us a 26-5 record.

Just to clarify further for non-nerds: say we play 10 games with a 60% chance of winning each. Our deterministic predicted record is 10-0. The probabilistic prediction is 6-4. Determinism gives poor forecasts, so it is likely that the RPIs in Warren Nolan's predicted team sheet will look quite different. It's still an interesting way to get a feel for what our resume may show in March.

Interesting stuff! I like the probabilistic prediction of 5 losses I think it's realistically probabilistic....lol! We should be favored in all games but 1 from here on out but that is not how it works.

[kskenyon4]

Originally Posted by waterhead

I don't recall our final RPI and Kenpom last year leading into the tourney but they were a lot closer than #7 and #34. They were probably closer to dead on around #12 or so within a few spots...again IDK.

Last year we ended at 12 RPI and 23 Kenpom.


Our RPI last year did this the final 16 weeks:
72
37
40
49
31
24
17
18
16
16
15
15
14
14
13
12

[justinhub2003]

Originally Posted by waterhead

I will continue to say this...RPI doesn't mean squat until the end when it starts to match the other metrics. But in the end it means a lot to the committee but they don't really evaluate this stuff until the end anyway.

Up until the last month or so Kenpom is better to look at while RPI is getting narrowed down.

Yea RPI is junk in that if your a terrible mid major school but play 13 tough non conference games, you get rewarded.

But those teams always drop way down once They get well into conference play.

Kenpom is just the best. I actually don’t hate BPI as well.

[sedziobs]

Today's RPI impact games:

Wyoming at New Mexico - Currently RPI 51, just barely into Group 2. 49% chance to win.

Miss St at Florida - Florida RPI 49 (Group 1), MS St RPI 65 (Group 2). Still lots of time left to determine where these teams fit in the SEC.

[waterhead]

Originally Posted by kskenyon4

Last year we ended at 12 RPI and 23 Kenpom.


Our RPI last year did this the final 16 weeks:
72
37
40
49
31
24
17
18
16
16
15
15
14
14
13
12

I think we were well above 100 just a couple weeks ago. RPI is junk until late season. We have been Kenpom between #5 and #15 I think the whole time. Buffalo is still our best win in RPI...lol! I mean...go Buffalo...but RPI isn't a metric to use until the last month of the season. For us with our schedule severely back loaded...we might as well not look at it until the committee does.

[sedziobs]

Our RPI is irrelevant. It's our opponents' RPIs that matter, more specifically what group they fall into. I think it's definitely worth keeping track of now.