Rankings next week - lots of ranked teams lost..

[Phil]

Humans programmed the computer. The code reflects the programmer's bias.

Oh please.

Obviously a tiny smidgen of truth (and in theory potentially significant — don't get me started on climate models) but you can't just throw out this with our providing at least some hint as to how this would be done.

Do you think the computer code literally keeps track of mid-majors and downrates them? That would be simply out and out fraud. If that's not what you mean what do you mean?

 

[Phil]

I’m just predicting the AP Top 25, which is what a media panel of 32(?) collectively decide. No personal bias used (other than giving a shoutout to Minnesota at #25).

Belmont was ranked #24 last week and unlike many teams ranked past the Top 13, they are at least winning the games they are supposed to.

Sorry, I do see that you were predicting the AP top 25 not your own version of what should be the top 25.

Gen.erally, I think the AP writers often do a decent job but I'm totally perplexed at why they have Belmont ranked so high.

I don't have the data but I wonder if there's ever been that large a disparity between a human poll and a computer poll.

 

[LoTrader]

I find the Massey Ratings to be most in line with my thinking of where teams should be ranked.

 

[DefenseBB]

I would say 2nd and 3rd best players (Watts and Walker).... I still say Paris Kea is their top player. Bailey will be the ACC rookie of the year..... they could compete for a NCAA berth next season.... as much as that pains me to say.

??? Isn't Sylvia Hatchell still the coach? I will take the under on that bet...

Side note, they also have a very real possibility of losing out the remaining games this year...can you please do your part in it???

 

[LoTrader]

Charlie Creme's bracket? This guy is ridiculous. He has Tennessee as a #3, Georgia as #4, NC State as #4, Maryland as #4. The only team in that group that I can argue should be in top 16 is Georgia. My goodness, this guy continues to unimpress.

 

[Jim]

Do you think the computer code literally keeps track of mid-majors and downrates them? That would be simply out and out fraud. If that's not what you mean what do you mean?

In order for the computer to come up with a single ranking number, various aspects about the team's play have to be encapsulated as numbers and those set of metrics are weighted to derive the final ranking. The choice of what to measure and how, along with the weights for each, is chosen by humans. Moreover, computer models assume transitivity (i.e, if A beats B, and B beats C, the A should beat C) which might not hold because of difficult to quantify metrics (i.e., team A has height but loses to a quick, pressing team, like C, or B is a good 3-pt shooting team at home but can't hit the broad side of a barn on the road.). If those metrics aren't in the model, they aren't reflected in the ranking and it was a human deciding what is in and what is not.

 

[Plebe]

Charlie Creme's bracket? This guy is ridiculous. He has Tennessee as a #3, Georgia as #4, NC State as #4, Maryland as #4. The only team in that group that I can argue should be in top 16 is Georgia. My goodness, this guy continues to unimpress.

First of all, he's only predicting what the committee will do, and for that he's largely basing himself off the two reveals so far.

But who are the 3 teams that you feel should be placed ahead of Tennessee, Maryland, and NC State?

 

[LoTrader]

First of all, he's only predicting what the committee will do, and for that he's largely basing himself off the two reveals so far.

But who are the 3 teams that you feel should be placed ahead of Tennessee, Maryland, and NC State?

Minnesota instead of Maryland, Oregon State instead of Tennessee, and Duke or LSU, as toss up with NC State.

 

[LoTrader]

First of all, he's only predicting what the committee will do, and for that he's largely basing himself off the two reveals so far.

But who are the 3 teams that you feel should be placed ahead of Tennessee, Maryland, and NC State?
Edit to add: In fairness, Charlie's last bracket came out before Maryland's losses to Purdue and Minnesota.

Question: is the next reveal tonight?

 

[Plebe]

Question: is the next reveal tonight?

Oops, I wrote that last "edit" before I saw today's bracket from Charlie.

Yes, the next and final reveal is tonight.

 

[Phil]

Charlie Creme's bracket? This guy is ridiculous. He has Tennessee as a #3, Georgia as #4, NC State as #4, Maryland as #4. The only team in that group that I can argue should be in top 16 is Georgia. My goodness, this guy continues to unimpress.

Help me out here.

In one post you say you like Massey (I agree).

Massey has TN at #12, which is a three seed.

Then you say Creme is "ridiculous" and you use as evidence that he has TN as a 3 seed.

Are there two different people using your login?

Edit to add:
Massey has MD at #16. A four seed. You say Creme is "ridiculous" for putting MD on the 4 line.

Which is it?

 

[Plebe]

Minnesota instead of Maryland, Oregon State instead of Tennessee, and Duke or LSU, as toss up with NC State.

Putting Minnesota in the top 16 right now would be a severe overreaction. Minnesota's resume right now is more like that of an 8 or 9 seed.

Oregon State gave itself a big boost with the win over UCLA, but their resume still isn't top-16 quality. Their RPI is just too low.

There is no way that Duke or LSU has a better resume than Tennessee. That's just not even close to being true.

 

[Phil]

In order for the computer to come up with a single ranking number, various aspects about the team's play have to be encapsulated as numbers and those set of metrics are weighted to derive the final ranking. The choice of what to measure and how, along with the weights for each, is chosen by humans. Moreover, computer models assume transitivity (i.e, if A beats B, and B beats C, the A should beat C) which might not hold because of difficult to quantify metrics (i.e., team A has height but loses to a quick, pressing team, like C, or B is a good 3-pt shooting team at home but can't hit the broad side of a barn on the road.). If those metrics aren't in the model, they aren't reflected in the ranking and it was a human deciding what is in and what is not.

That's not an answer

As an aside, computer models do not build transitivity into their models.

What about your answer has anything to do with why the models are biased against mid-majors?

As another aside, I do know something about models. I have built computer models of many thing, including Hepatitis C, breast implants, Y2K, hurricane incidence angles, auto warranty claims, nuclear terrorists attacks and much more. I know a lot about computer modeling. The next model I see that assumes transitivity will be the first one I've ever seen. And I'll poke fun at it if it actually exists.

 

[Phil]

Putting Minnesota in the top 16 right now would be a severe overreaction. Minnesota's resume right now is more like that of an 8 or 9 seed.

I agree. Massey has them at 29, which sounds about right. Solid 8 seed.

 

[LoTrader]

Putting Minnesota in the top 16 right now would be a severe overreaction. Minnesota's resume right now is more like that of an 8 or 9 seed.

Oregon State gave itself a big boost with the win over UCLA, but their resume still isn't top-16 quality. Their RPI is just too low.

There is no way that Duke or LSU has a better resume than Tennessee. That's just not even close to being true.

I understand where you are coming from....but RPI is a joke, because it is based off of the preseason rankings which are never accurate. It may turn out as Charlie Creme has predicted, but it doesn't pass the smell test. Tennessee doesn't pass the smell test. Their play stinks. Tennessee (and Maryland) are on a downward slope, and OSU is on an upward slope. If nobody had a "home court" advantage during the 1st 2 playoff games, I wouldn't care too much about the seeding. It would be nice if there was consideration given to how teams are playing at the time of the reveal.

 

[Plebe]

I understand where you are coming from....but RPI is a joke, because it is based off of the preseason rankings which are never accurate. It may turn out as Charlie Creme has predicted, but it doesn't pass the smell test. Tennessee doesn't pass the smell test. Their play stinks. Tennessee (and Maryland) are on a downward slope, and OSU is on an upward slope. If nobody had a "home court" advantage during the 1st 2 playoff games, I wouldn't care too much about the seeding. It would be nice if there was consideration given to how teams are playing at the time of the reveal.

RPI has absolutely nothing to do with preseason rankings. Every team starts with a clean slate.

Unfortunately, Oregon State did nothing in the nonconference, other than lose to two good teams and beat a whole bunch of horrible teams.

 

[Shades]

My attempt to predict the AP Top 25
1. Connecticut
2. Mississippi St
3. Baylor
4. Notre Dame
5. Louisville
6. Texas
7. South Carolina
8. Oregon
9. UCLA
10. Florida St
11. Missouri
12. Oregon St
13. Ohio St
14. Tennessee
15. Maryland
16. Stanford
17. Texas A&M
18. South Florida
19. Duke
20. Belmont
21. North Carolina St
22. Georgia
23. Oklahoma St
24. Green Bay
25. Louisiana St (or Minnesota?)

Out: Michigan

Actual AP Top 25
1. Connecticut
2. Mississippi St
3. Baylor
4. Louisville
5. Notre Dame
6. Texas
7. South Carolina
8. Oregon
9. Florida St
10. UCLA
11. Missouri
12. Oregon St
13. Maryland
14. Ohio St
15. Tennessee
16. Stanford
17. Texas A&M
18. South Florida
19. Georgia
20. Duke
21. North Carolina St
22. Green Bay
23. Belmont
24. Louisiana St
25. Oklahoma St
———————————
26. Minnesota
27. Dayton
28. Mercer
29. Miami
30. Arizona St

Out: Michigan

 

[LoTrader]

RPI has absolutely nothing to do with preseason rankings. Every team starts with a clean slate.

Unfortunately, Oregon State did nothing in the nonconference, other than lose to two good teams and beat a whole bunch of horrible teams.

You are correct, I was wrong in my assertion....RPI does not have anything to do with preseason rankings. My new conjecture is RPI has far too much influence in the current ranking system, and, does not take into consideration how a team is trending.

Re: OSU's non conference record, OSU lost to ND by 5 (at home) in mid-November, and to Duke by 7 (at Duke) in late November. I consider Notre Dame to be a pretty solid team, and having been in attendance at that game, Notre Dame won due to their experience. OSU led nearly the entire game until the Notre Dame guards made some great plays. I hope OSU gets another crack at Notre Dame. The Duke game was different. OSU's defense was dodgy, and Duke played better. Again, while you want to diminish OSU's out of conference record, and specifically these two losses, they were early in the season, and, as I submit now, OSU is playing the best basketball they have played to date. Not that a team's overall record shouldn't be a major factor in their ranking, in addition to the teams they have played to date...I don't think these rating systems (other than Massey) give enough consideration to how a team is trending.

 

[Jim]

That's not an answer

As an aside, computer models do not build transitivity into their models.

What about your answer has anything to do with why the models are biased against mid-majors?

As another aside, I do know something about models. I have built computer models of many thing, including Hepatitis C, breast implants, Y2K, hurricane incidence angles, auto warranty claims, nuclear terrorists attacks and much more. I know a lot about computer modeling. The next model I see that assumes transitivity will be the first one I've ever seen. And I'll poke fun at it if it actually exists.

RPI and SOS indices are based on transitivity -- record against opponents and the opponents record other opponents, and so on, which is transitivity.

I know a lot about computer models too. You aren't the only one who does it for a living. None of the models you mention involve ranking competitors. Ranking models need to incorporate transitivity to develop ranking order.

 

[LoTrader]

RPI and SOS indices are based on transitivity -- record against opponents and the opponents record other opponents, and so on, which is transitivity.

I know a lot about computer models too. You aren't the only one who does it for a living. None of the models you mention involve ranking competitors. Ranking models need to incorporate transitivity to develop ranking order.

What ranking models to you consider to be the best available (at a given moment in time)?

 

[EricLA]

I want Tennessee to drop to a 4 seed and be put in UCONN's bracket. That would be fantabulous. Or, if Baylor stays a 2 seed (which is unlikely IMHO), then they can put Tenn as Baylor's 3-seed...

 

[DefenseBB]

For the most part, today's AP Ranking is about right as I too subscribe more to Massey and my eye test on recent play. Some teams are not trending well but their Massey ranking warrants where they are at the moment. Tennessee, Maryland, Duke, ASU all have tough games left and will probably drop more. I honestly think only 1 Big 10 should be in the top 16 and tOSU has the easiest schedule left.

Tonight's reveal will be interesting for Baylor, ND, Louisville and Oregon as all want to avoid the #2 seed in Albany.

Some teams are trending very well and will get good seeds in their Conference Tourney's-Miami, LSU, Minnesota, Georgia will improve their ranking this week.
Some curious comments I feel:

• I don't see USF as a top 20 team. They will have 1 more loss coming and are at best a #6 seed.
• I think SC at #7 overall is a reach as well. They should a be #3 at best (9-11) and their games against LSU/Tenn this week may help or hurt. I get that as the 2nd best team in the SEC, the committee will over-seed them but I take Oregon, FSU and UCLA ahead of them.
• Miami is coming on and should get the #4 seed in the ACC along with a double bye.

I am sad that UCF lost in OT yesterday as that would have helped their NCAA cause. I think they are "dead man walking now" regardless of their AAC Tourney and are NIT bound unless, of course, they win the AAC...

Our conference is still feeble no matter how we slice it. Hopefully our subs continue their evolution from yesterday and showing well against USF would be a good boost.

 

[Plebe]

You are correct, I was wrong in my assertion....RPI does not have anything to do with preseason rankings. My new conjecture is RPI has far too much influence in the current ranking system, and, does not take into consideration how a team is trending.

Re: OSU's non conference record, OSU lost to ND by 5 (at home) in mid-November, and to Duke by 7 (at Duke) in late November. I consider Notre Dame to be a pretty solid team, and having been in attendance at that game, Notre Dame won due to their experience. OSU led nearly the entire game until the Notre Dame guards made some great plays. I hope OSU gets another crack at Notre Dame. The Duke game was different. OSU's defense was dodgy, and Duke played better. Again, while you want to diminish OSU's out of conference record, and specifically these two losses, they were early in the season, and, as I submit now, OSU is playing the best basketball they have played to date. Not that a team's overall record shouldn't be a major factor in their ranking, in addition to the teams they have played to date...I don't think these rating systems (other than Massey) give enough consideration to how a team is trending.

The committee does (or at least can) consider "early competition versus late competition": Women's Basketball Selections 101 - Selections

But that doesn't mean ignoring the first half of the season either. They judge the entire body of work.

For Oregon State that nonconference SOS is an albatross: #253. That spells ouch. The only team in the top 60 with a lower NCSOS is ... Minnesota.

 

[Plebe]

RPI and SOS indices are based on transitivity -- record against opponents and the opponents record other opponents, and so on, which is transitivity.

That isn't what transitivity means.

 

[LoTrader]

Help me out here.

In one post you say you like Massey (I agree).

Massey has TN at #12, which is a three seed.

Then you say Creme is "ridiculous" and you use as evidence that he has TN as a 3 seed.

Are there two different people using your login?

Edit to add:
Massey has MD at #16. A four seed. You say Creme is "ridiculous" for putting MD on the 4 line.

Which is it?

Phil, you've got me there. There are not two different people using my login....that was funny! To be precise, I did say that I agree with the Massey ranking system more than any other (but not 100%). If I had to go with one of the ranking systems, and live with the results of that system, it would be Massey.