This mornings Pomeroy ratings

[Billikenswin]

We moved up to 21. Defense now sits at #2 in the country. Offense dropped from 100 to 122. Fix some of the problems on offense and this can be a really special season.

I know that some like the Pomeroy system and some go to the RPI sites but I think the best site is Massey rankings which compiles and averages the results of several dozen ratings systems and predicts the results of all games based upon actual results going forward. They have us at 29 which seems about right to me. That will move up or down with the results of future games of ours and of our opponents with adjustments constantly for the variations in the compilation of ratings and rankings. It seems solid to me and is often very close to our man --the Wiz--in terms of predictions.

[kshoe]

I feel like that's just the nature of how college basketball games go and that in a model like Pomeroy's those intricacies would level out over the course of a season.

I agree that it should level out over the season. That being said, I think we've benefited by it more than not this season as we've had quite a few close games turn into routes in the final 6 minutes or so. We also had a strong comeback against UW that turned a double digit deficit into only a 6 point loss. The ones that went against us were Wichita St. (it was obviously closer than the 5 point margin of victory) and Valpo (we had a large lead most of the game). Maybe these late game swings account for why Pomeroy has us at 21, a few notches above where most Billiken fans would reasonably believe we should be?

[JettFlight5]

I saw Pomeroy and 122 and thought we were going back to all the old memes.

[Old guy]

The problem with all of these ratings is that there is no clear disclosure as to how they are determined. Since you (plural means all of us) do not know how they get to their numbers all you can do is speculate about it. This is why a compilation of various systems may be the better solution.

[JMM28]

Pomeroy has a detailed breakdown of his numbers on his website. RPIforecast has a detailed breakdown of their numbers as well for every single team.

[kshoe]

Pomeroy has a detailed breakdown of his numbers on his website. RPIforecast has a detailed breakdown of their numbers as well for every single team.

JMM28 1 Old Guy 0

[slufanskip]

I doubt anybody that watched that game would describe it as an 8 point victory. It was a super tight game with an inflated margin of victory based on fts and fouling at the end.

Actually it was a slow start and then a dominating 2nd half. We won the 2nd half by 14. We were clearly the better team and I believe the margin of victory reflects the game. We caught up the halftime deficit of 6 in the 1st 10 or so minutes of the 2nd half and then won the next 10 by 8. We were up give or take 7 points before they started fouling intentionally. We really didn't stretch the margin at the end

[Old guy]

JMM28 1 Old Guy 0

You got that right

[kshoe]

Actually it was a slow start and then a dominating 2nd half. We won the 2nd half by 14. We were clearly the better team and I believe the margin of victory reflects the game. We caught up the halftime deficit of 6 in the 1st 10 or so minutes of the 2nd half and then won the next 10 by 8. We were up give or take 7 points before they started fouling intentionally. We really didn't stretch the margin at the end

We were trailing from basically the 8 minute mark in the first half to the 3:30 minute mark in the second half (23 minutes of the 40 minute game) when we tied it up. If that is a game that we "dominated" and an 8 point margin of victory than so be it.

Maybe I'll put it differently, an 8 point spread on a game suggests a game a team should win ~80% of the time. With 3:30 to go and the game tied I sure didn't feel like this was a game we'd win 80% of the time.

[DD3]

I feel like that's just the nature of how college basketball games go and that in a model like Pomeroy's those intricacies would level out over the course of a season.

Correct. If you're going to devalue an "inflated" win that results from end-of-game fouling, then you have to give added weight to those wins that were closer than expected due to sitting your starters for the last five minutes. And you'll have to give some added weight to victories where your best player had the flu. And you'll have to discount for losses that occurred on the road when your players were in the middle of final exams. We start judging games based on aesthetics and not results. Slippery slope.

Like NH said--it's likely that all of these things even out over the course of a long season.

[slufanskip]

We were trailing from basically the 8 minute mark in the first half to the 3:30 minute mark in the second half (23 minutes of the 40 minute game) when we tied it up. If that is a game that we "dominated" and an 8 point margin of victory than so be it.

Maybe I'll put it differently, an 8 point spread on a game suggests a game a team should win ~80% of the time. With 3:30 to go and the game tied I sure didn't feel like this was a game we'd win 80% of the time.

You're right my bad. Just on memory and watching pretty late last night I was thinking we stretched it out earlier.

[waldo027]

We were trailing from basically the 8 minute mark in the first half to the 3:30 minute mark in the second half (23 minutes of the 40 minute game) when we tied it up. If that is a game that we "dominated" and an 8 point margin of victory than so be it.

Maybe I'll put it differently, an 8 point spread on a game suggests a game a team should win ~80% of the time. With 3:30 to go and the game tied I sure didn't feel like this was a game we'd win 80% of the time.

I don't think anyone said we dominated the entire game. But the score was 53-44 with about a minute left. We went on a 16-5 run in the final 5 minutes. We did not just hit a bunch of meaningless free throws in the final seconds to extend the lead. A 3 point spread implies it should come down to the final possession or two, but we had this game comfortably in hand by that point.

[slufanskip]

I don't think anyone said we dominated the entire game. But the score was 53-44 with about a minute left. We went on a 16-5 run in the final 5 minutes. We did not just hit a bunch of meaningless free throws in the final seconds to extend the lead. A 3 point spread implies it should come down to the final possession or two, but we had this game comfortably in hand by that point.

He was partially right regarding my post. I implied we had it in hand earlier than we did. Though we didn't increase the margin due to meaningless f/t's. To be honest, I was never really nervous. I felt like we were the better team and would win even with 5 minutes left and it still being within a couple of points or so.

[The Wiz]

These things do even out in the long run....that's why Nov is always a tough month to forecast...limited data....more data=more accurate.

I have the Bills at #30....earlier in the season I was ahead of Pomeroy now he has moved ahead of me at least in regards to the Bills ranking.

With 1:55 left in the game, it was Bills by 5. Even when I miss on a spread, many times you will see the exact spread in the final 5 minutes. As to what happens once the game is decided, as mentioned previously, these things even out.

[Old guy]

Withdrawn the format does not copy properly in the post.

[MB73]

You were critical of the computer rankings all year last year. Whether it was Pomeroy, Sagarin or the Wiz, you constantly excoriated them for being wrong. Then, we end up in the top 20 teams at year end. Maybe it was you that was completely, totally, and unambiguously wrong the entire year!

For the record, I never criticized Sagarrin, though he is not flawless. But KenPom and The Wiz, yes, I did. I prefer RPI though it is early for all of these models.

Last year we finished with a flurry, 16-2, and earned it! Beat some good teams.

At the time I was criticizing KenPom and The Wiz, we had not beaten a top 50 team and were about 12-5 and I said we were 40th or so. They had us 12th or 13th.

Same thing this year. Sure, if we go 16-2 again then we are in top 20 or so.

And do not start on the predictive model bs. We got hot, did the job, good for us.

[jimbofive]

Withdrawn the format does not copy properly in the post.

if you don't have anything nice to say, say oregon cheerleaders.

[almaman]

Correct. If you're going to devalue an "inflated" win that results from end-of-game fouling, then you have to give added weight to those wins that were closer than expected due to sitting your starters for the last five minutes. And you'll have to give some added weight to victories where your best player had the flu. And you'll have to discount for losses that occurred on the road when your players were in the middle of final exams. We start judging games based on aesthetics and not results. Slippery slope.

Like NH said--it's likely that all of these things even out over the course of a long season.

yep

[The Wiz]

For the record, I never criticized Sagarrin, though he is not flawless. But KenPom and The Wiz, yes, I did. I prefer RPI though it is early for all of these models.

Last year we finished with a flurry, 16-2, and earned it! Beat some good teams.

At the time I was criticizing KenPom and The Wiz, we had not beaten a top 50 team and were about 12-5 and I said we were 40th or so. They had us 12th or 13th.

Same thing this year. Sure, if we go 16-2 again then we are in top 20 or so.

And do not start on the predictive model bs. We got hot, did the job, good for us.

I think I finally get it....

1. Data analysis...Using a weather analogy....Dave Murray is the Fox 2 meteoroligist with a long career that stretches back to forecasting on national TV.... He, with his foo foo computers and electronic weather maps to make prognostications. Ah, but we are not fooled anymore since we have been enlightened. We now know what Dave really does....He calls the weather bureau each day. Anyone can do it.. Anyone that believes otherwise is being taken in. Has anyone checked Dave's phone records.....Reality

2) Predictive models....Again, being enlightened I now realize these do not exist. Of course, it was pointed out to me that instead of trying to predict things that are most likely to happen (tip o' the cap to Descartes) it would better to predict things that are unlikely....hmmm ...maybe such as predicting the Bills rankings in the teens a couple of years ago while others had us in the 60's....and again last year when I had us in the teens early on ...it was pointed out this was poopy. When these things actually came to be ...the story is ...The Bills got hot...I "fell into it " (the prediction) Seems to me you can't have it both ways.. Or maybe in reality, you can have it both ways.

Poor KenPom ...he is not even here to defend himself.

So in summary....Data analysis is just calling up someone (Vegas or the weather bureau) who has data and repeating their stuff....if the final figures come out close ...it has nothing to do with probability or statisical analysis...it just proves Dave's a thief and a sham.

Further, predictive models are just wild shots in the dark that are like winning the lotto...you just "fall into it" when the Bills accidentally win a bunch in a row.

Well that about wraps it up.....wait there is one last loose end.

What about those predictions of the exact number of 3 point shots made by the Bills opponents... That just doesn't seem to fit in anywhere....hmmm....I know what it is

Oh, oh, oh

It's magic you know

Never believe it's not so

Pilot (1974)

[AlumniFan]

I think I finally get it....

1. Data analysis...Using a weather analogy....Dave Murray is the Fox 2 meteoroligist with a long career that stretches back to forecasting on national TV.... He, with his foo foo computers and electronic weather maps to make prognostications. Ah, but we are not fooled anymore since we have been enlightened. We now know what Dave really does....He calls the weather bureau each day. Anyone can do it.. Anyone that believes otherwise is being taken in. Has anyone checked Dave's phone records.....Reality

2) Predictive models....Again, being enlightened I now realize these do not exist. Of course, it was pointed out to me that instead of trying to predict things that are most likely to happen (tip o' the cap to Descartes) it would better to predict things that are unlikely....hmmm ...maybe such as predicting the Bills rankings in the teens a couple of years ago while others had us in the 60's....and again last year when I had us in the teens early on ...it was pointed out this was poopy. When these things actually came to be ...the story is ...The Bills got hot...I "fell into it " (the prediction) Seems to me you can't have it both ways.. Or maybe in reality, you can have it both ways.

Poor KenPom ...he is not even here to defend himself.

So in summary....Data analysis is just calling up someone (Vegas or the weather bureau) who has data and repeating their stuff....if the final figures come out close ...it has nothing to do with probability or statisical analysis...it just proves Dave's a thief and a sham.

Further, predictive models are just wild shots in the dark that are like winning the lotto...you just "fall into it" when the Bills accidentally win a bunch in a row.

Well that about wraps it up.....wait there is one last loose end.

What about those predictions of the exact number of 3 point shots made by the Bills opponents... That just doesn't seem to fit in anywhere....hmmm....I know what it is

Oh, oh, oh

It's magic you know

Never believe it's not so

Pilot (1974)

Wiz - when are you going to learn that a man who writes in Kiplinger style (whatever the hell that is) is far superior to the rest of us. Math is foo foo anyways. Next thing you know, you'll be trying to convince us the world is round!!!

[MB73]

I know the difference between "data analysis" and "predictive models" and "bullsh*t." KenPom with us at 21 is bullsh*t.

SLU is about 35th, 40th now, give or take, we haven't beaten anybody, but sure if we finish like last year, a 16-2 type run, beat Dayton, VCU,

U Mass, win the A-10 & A-10 Tournament, win two NCAA games, then sure, we will be Top 15-20.

Predictive model? Did KenPom and / or The Wiz predict Louisville to win it all last year? Michigan, Wichita State, Syracuse to get so far?

(bizzleboy, look up his rankings last year at this time)

[Quality Is Job 1]

MB is a glutton for punishment.

[The Wiz]

I know the difference between "data analysis" and "predictive models" and "bullsh*t." KenPom with us at 21 is bullsh*t.

SLU is about 35th, 40th now, give or take, we haven't beaten anybody, but sure if we finish like last year, a 16-2 type run, beat Dayton, VCU,

U Mass, win the A-10 & A-10 Tournament, win two NCAA games, then sure, we will be Top 15-20.

Predictive model? Did KenPom and / or The Wiz predict Louisville to win it all last year? Michigan, Wichita State, Syracuse to get so far?

(bizzleboy, look up his rankings last year at this time)

LOL....As a matter of fact, if you check the Billiken Board March Madness Bracketology Challenge held on this site last March, you will see in fact I picked Louisville to win it all.....Oops.

[JMM28]

MB - where does mom have SLU?

[Old guy]

if you don't have anything nice to say, say oregon cheerleaders.

Oregon cheerleaders