NCAA Bracketology - The Week of February 18th

[07Illini]

Originally Posted by Illwinsagain

Well, we are seeded the same as Wichita St., La Salle and UNC. Most years I would be happy being seeded the same as UNC, but not this year.

What a joke. UNC has exactly 1 win against an RPI top 50 team. Lunardi still has Kentucky as one of the first four out after getting totally dismantled by Tenn. His bracket is garbage.

[AHSIllini32]

Originally Posted by scubadunk

http://espn.go.com/mens-college-basketball/bracketology

10 seed on Lunardi's Bracketology. Playing 7 seed NC State.

Garbage

[AEX]

Originally Posted by 07Illini

What a joke. UNC has exactly 1 win against an RPI top 50 team. Lunardi still has Kentucky as one of the first four out after getting totally dismantled by Tenn. His bracket is garbage.

Completely agree. Also, Minnesota is an 8 seed, the same team that we have swapped wins with, who is tied with us in the B1G standings, and who has fewer marque wins. I'm also calling BS on the Big East getting more teams in.

This may not be the place, but it just kills me how people acknowledge how tough playing in the B1G is, how difficult it is to play 5 games in a row against ranked teams, but then fault teams for struggling. It's a joke.

[Botb9]

Originally Posted by AEX

Completely agree. Also, Minnesota is an 8 seed, the same team that we have swapped wins with, who is tied with us in the B1G standings, and who has fewer marque wins. I'm also calling BS on the Big East getting more teams in.

This may not be the place, but it just kills me how people acknowledge how tough playing in the B1G is, how difficult it is to play 5 games in a row against ranked teams, but then fault teams for struggling. It's a joke.

BPI supposedly takes teams and ranks them according to the result when they play an average game on a neutral floor. Illinois has not once played an "average game". It's not a good stat for Illinois to be rated by, but as I commented, to point out the bad only and to indicate that this is the reason they should be on the bubble is disingenuous.

[WesterveltVictoryCigar]

The BPI is interesting, not quite sure on the agenda, but parts of the explanation are interesting:

There are a number of small details that we have in our methodology to make it reflective of a résumé for a tournament team -- these are pretty technical and many people won't be interested, so we won't go into detail, but we think they improve how the tool works.

That's OK, I guess. Sagarin doesn't publish his formulas and others are probably equally secretive.

Ultimately, the College Basketball Power Index gives us a tool for rating teams that is more useful for ESPN than existing tools. It explains a team's wins and losses in a more detailed context and seemingly predicts future results as well as other tools, if not better. It also incorporates information about injuries and missed games that should be relevant for weighting the résumés of bubble teams on Selection Sunday. We feel confident that it will enhance coverage of college basketball for years to come.

Ah.....

In all seriousness, I'm sure it's a valid attempt to come up with a new tool, but it may need some tweaking. It's interesting that they attempt to account for players missing from a team, but that also seems to be a bit of a slippery slope. A replacement may play better than the missing player for that particular game, it would seem difficult to me to accurately reflect the impact.

From an Illini perspective, BPI is by far the outlier when you look at the compilation of ranking systems. It also seems to greatly diminish the performance of one of our big wins, Butler. Given their recent history in the tournament, maybe that's a compliment to us after all?

[IlliniInOK]

The west in Palm's bracket really is hilariously bad... a weak 1 seed in Gonzaga who has maybe 3 wins that would be considered good - Kansas State, Ok. State and Oklahoma, a weak 2 seed in Arizona who has 2 really good wins - Florida (at home) and Miami (FL) at home, a decent home win against San Diego State (a bubble team) and then a bunch of nothing... New Mexico is the 3 with their best wins being Cincinnati and UCONN, a couple of decent Big East teams that are overrated because they're in the Big East

[MA Illini]

That's OK, I guess. Sagarin doesn't publish his formulas and others are probably equally secretive.

This is an interesting point. I understand why ESPN wants it to be secretive, so they can be the only ones to publish it. However, hard to gain credibility when you don't publish what you are doing. ESPN is in a better position to publish what they are doing than anyone else. Hard for people to rip off ESPN when they are the Juggernaut vs. if some random blogger came up with BPI and then ESPN copied the formula and posted the ranking on ESPN.com

Sagarin is respected and thought to be agenda/conflict free. ESPN has not had the best of reputation about this lately and does have clear financial conflicts with is TV contracts (outside of the ever present media conflict of attracting attention vs. reporting the truth). However, just because conflicts exist doesn't mean there is manipulation occurring.

From an Illini perspective, BPI is by far the outlier when you look at the compilation of ranking systems. It also seems to greatly diminish the performance of one of our big wins, Butler. Given their recent history in the tournament, maybe that's a compliment to us after all?

Agreed, BPI is the clear outlier for the Illini. Most teams probably have one ranking that underrates/overrates them significantly. Not sure what you mean about the Butler win though. We had a BPI 98.9 for that game.

[IlliniFan95]

Don't worry, when we go 2-0 this week we'll hear more good reasons to downplay our wins in a weak Big Ten!

[WesterveltVictoryCigar]

Originally Posted by MA Illini

Not sure what you mean about the Butler win though. We had a BPI 98.9 for that game.

Sorry I was unclear here. BPI has Butler ranked well below where most other systems have them as well.

[IlliniRyan13]

Originally Posted by AHSIllini32

Garbage

Total garbage. Lunardi has also been horrific as a bracketologist past few years. So the worse he seeds us, the better I feel.

[MA Illini]

Originally Posted by WesterveltVictoryCigar

Sorry I was unclear here. BPI has Butler ranked well below where most other systems have them as well.

Gotcha. Although I think the way the BPI works, the BPI score for the game takes into account the strength of the opponent. So we had game BPI of 98.9 even thought it doesn't think Butler is that good. So would have gotten a higher score, 99+, if it thought Butler was better but it would not really affect our BPI ranking.

This does bring up an interesting point about the BPI thought. You could have a great ranking raking up wins over inferior competition. We had a 92.2 against E. Kentucky. So if we played E. Kentucky 26 times the same way, BPI would rank us the #1 in the country. E. Kentucky is ranked #106 by the BPI and 65th by RPI.

This brings up my point from earlier thought about SOS. E Kentucky is ranked 65th in the RPI and our average RPI ranking for opponents is around 100. A schedule of 26 teams ranked 65th in RPI would give us the toughest schedule in the nation by a wide margin given how the SOS metrics work. So if we went 19-8 against E. Kentucky, we would be ranked higher than where were are now. I dont think anyone would think we were a better team than we are now by doing 19-8 against E. Kentucky, but the numbers would suggest other wise.

[HOOPDOG]

You can't have three teams from the Big Ten all in the East

[DaytonIllini]

Originally Posted by blmillini

And, if it were done correctly that would be a fine argument. It is obvious to me that the SOS component of that index is total garbage.

When you take a class and are attempting to attain any grade you expect the tests that comprise that grade to be relevant to the topic. If the test that caused your grade to slip was inconsistent with what any reasonable professor would have taught in the class, then that professor needs to be fired and needs to change... much like the BPI.

In fairness a student that had 7 A's and 4 F's could hardly expect to get anything other than a C. In their system narrow wins against bad teams hurt you. You'd have to blow out several great teams to offset a couple of times you got spanked by bad teams.

Personally I think margin of victory is silly though. A coach losing by 4 doesn't play to keep the game from getting out of hand. He plays to win and starts fouling. To him, losing by 4 or 20 is the same. But that close loss that turns into a blowout when your opponent makes his FT's hurts you. Likewise when we are up 4 with the clock winding down we don't run our offense and launch a last second 3 in order to run up the margin of victory. That component of analysis should not be used when ranking teams in my opinion.

[MA Illini]

Originally Posted by DaytonIllini

In fairness a student that had 7 A's and 4 F's could hardly expect to get anything other than a C. In their system narrow wins against bad teams hurt you. You'd have to blow out several great teams to offset a couple of times you got spanked by bad teams.

Personally I think margin of victory is silly though. A coach losing by 4 doesn't play to keep the game from getting out of hand. He plays to win and starts fouling. To him, losing by 4 or 20 is the same. But that close loss that turns into a blowout when your opponent makes his FT's hurts you. Likewise when we are up 4 with the clock winding down we don't run our offense and launch a last second 3 in order to run up the margin of victory. That component of analysis should not be used when ranking teams in my opinion.

The definition of "narrow" victory is also a subjective one and related to your point about margin of victory. I agree with the margin of victory argument, there are so many factors at the end of the game where the incentive to maximize(minimize) margin of victory (defeat) is not aligned with the incentive to win the game. It has been shown to be generally predictive of the outcome of rematches (see Ken Pom blog post from a few weeks back) and I think everyone agrees it is important, but it has clear flaws.

Take our game against Chaminade. They are DII so they automatically get assigned a rank of the worst team in DI by the BPI (they dont count in the RPI). Our raw BPI for the game was 95.0 and we won by 23 pts, clearly not a close game. But the raw BPI is adjusted down to 56.4 due to the opponent strength. So our BPI went down because we beat Chaminade by 23 pts! If would have been better to not play them at all. This happens with RPI also when you play teams ranked really low. This is a clear flaw in the rating systems, regardless of whether or not margin of victory is accounted for or not. How much should we have beat Chaminade by in order to maintain our BPI, 40 pts? If you go back and read some of my earlier posts, it all relates to the problem of trying to rank all 347 DI teams but using the ranking to determine the top 50 teams. There really isnt a difference between a top 50 team playing a team ranked #300 vs. #347, the top 50 team will win 95% percent of the time. But that difference in ranking between 300 and 350 is treated the same as the difference between playing #1 vs. playing #50, which has a much larger impact on the expected outcome for a top 50 team. For the average team, ranked #175, the difference between #300 and #350 maybe about the same as #50 vs. #1, but we are using the RPI/BPI/whatever ranking you want to distinquish between top 50 team, not teams 1-347.

[troyk]

Originally Posted by DaytonIllini

In fairness a student that had 7 A's and 4 F's could hardly expect to get anything other than a C. In their system narrow wins against bad teams hurt you. You'd have to blow out several great teams to offset a couple of times you got spanked by bad teams.

Personally I think margin of victory is silly though. A coach losing by 4 doesn't play to keep the game from getting out of hand. He plays to win and starts fouling. To him, losing by 4 or 20 is the same. But that close loss that turns into a blowout when your opponent makes his FT's hurts you. Likewise when we are up 4 with the clock winding down we don't run our offense and launch a last second 3 in order to run up the margin of victory. That component of analysis should not be used when ranking teams in my opinion.

Or when a team is down 10 and heaves up 2 quick 3s at the end to lose by 4. The outcome was never in doubt but instead of it being a double-digit win, it is a "bad" win by BPI metrics.

[troyk]

Originally Posted by MA Illini

The definition of "narrow" victory is also a subjective one and related to your point about margin of victory. I agree with the margin of victory argument, there are so many factors at the end of the game where the incentive to maximize(minimize) margin of victory (defeat) is not aligned with the incentive to win the game. It has been shown to be generally predictive of the outcome of rematches (see Ken Pom blog post from a few weeks back) and I think everyone agrees it is important, but it has clear flaws.

Take our game against Chaminade. They are DII so they automatically get assigned a rank of the worst team in DI by the BPI (they dont count in the RPI). Our raw BPI for the game was 95.0 and we won by 23 pts, clearly not a close game. But the raw BPI is adjusted down to 56.4 due to the opponent strength. So our BPI went down because we beat Chaminade by 23 pts! If would have been better to not play them at all. This happens with RPI also when you play teams ranked really low. This is a clear flaw in the rating systems, regardless of whether or not margin of victory is accounted for or not. How much should we have beat Chaminade by in order to maintain our BPI, 40 pts? If you go back and read some of my earlier posts, it all relates to the problem of trying to rank all 347 DI teams but using the ranking to determine the top 50 teams. There really isnt a difference between a top 50 team playing a team ranked #300 vs. #347, the top 50 team will win 95% percent of the time. But that difference in ranking between 300 and 350 is treated the same as the difference between playing #1 vs. playing #50, which has a much larger impact on the expected outcome for a top 50 team. For the average team, ranked #175, the difference between #300 and #350 maybe about the same as #50 vs. #1, but we are using the RPI/BPI/whatever ranking you want to distinquish between top 50 team, not teams 1-347.

They probably should assign the same RPI rank to teams below 250 or even 200.

[MA Illini]

Originally Posted by troyk

They probably should assign the same RPI rank to teams below 250 or even 200.

That could be part of it, although you are by default losing some information by truncating. I would be inclined to just de-emphasize results against RPI 200+ teams. They simply don't tell us much about a top 50 team. They certainly don't tell us near as much about games vs other top 50 or top 100 teams. The results should count for something and you have be careful with this approach also because so can't completely ignore things like TCU beating Kansas. But the selection committee already acknowledges that its the games against good teams that matter the most, so why not build it into the RPI or other rankings?

Just take the current RPI or BPI and apply a weighting factor based on the ranking. It should not be linear, but instead something higher order that counts games against 200+ opponents only some percentage (say 50%) of what results against top 50 opponents count for. This would be really easy to do and study how this affect the prediction of tournament results.

[solbur2]

IDC about stupid lunardi, or the dumb BPI. We are in the NCAA tournament as far as I'm concerned. Top wins Top RPI Top SOS.

[solbur2]

Lundari responding to a question about us being a 10..... "The Illini are actually a No. 9 seed on the S-Curve, but dropped a line to avoid early-round conflicts with other Big Ten entries. They are rising about a line a week at this point and should continue to do so as the conference record improves"

[Calvin]

Originally Posted by scubadunk

ESPN keeps talking up their BPI system that is severely flawed. A 11 lost Stanford team and an 11 loss Dayton teams are rated ahead of us. Now they post this article on their blog. ESPN is severely flawed.

ESPN can do whatever it wants. They are the microsoft* of sports. Maybe one day the conference networks will figure out how to bypass ESPN and still make as much money, but that day seems to be a long ways away. Btw, I agree completely with your point. With so many proprietary metrics out there, I suppose they figured they'd give it a try and see if it catches on. Probably had an intern come up with it.

*or insert most salient near-monopoly

[IlliniWC]

I wouldn't get too worked up about BPI - it seems like it's new, and they're still likely tweaking it.

[Calvin]

Originally Posted by solbur2

IDC about stupid lunardi, or the dumb BPI. We are in the NCAA tournament as far as I'm concerned. Top wins Top RPI Top SOS.

Yep. The bracket matrix has us averaging around a 7 seed, and EVERY single bracketologist has us in. Historically, bad losses aren't a big deal --pretty much every team has one or two. The problem is when you pile up bad losses. In the scheme of things, our schedule was too tough to have many. We certainly have more than we'd like, but we played some really tough teams, and our slump didn't span too many bad teams.

Another really ignorant thing is claiming wins aren't really worth much in the eyes of the committee because a team didn't win by enough points. Dumb logic, and a sign the writer is stretching things.

This year in particular should be good for us because the committee has acknowledged imbalances in schedules, and we're in a brutal conference.

[scubadunk]

Originally Posted by IlliniWC

I wouldn't get too worked up about BPI - it seems like it's new, and they're still likely tweaking it.

From how they describe the BPI it sounds like it needs a complete overhaul. I think its just another way for ESPN to try and ingratiate themselves in the NCAA selection process for the tourney. Apparently logic and all the other systems out there just arent good enough for ESPN. I would like to see their justification for having a Stanford or a Dayton ahead of Illinois or even Illinois State being one spot lower than Illinois despite having no big resume building wins. I really think ESPN just wants to see this as part of the equation on selection sunday just as they have with Bracketology.

[Botb9]

Originally Posted by scubadunk

From how they describe the BPI it sounds like it needs a complete overhaul. I think its just another way for ESPN to try and ingratiate themselves in the NCAA selection process for the tourney. Apparently logic and all the other systems out there just arent good enough for ESPN. I would like to see their justification for having a Stanford or a Dayton ahead of Illinois or even Illinois State being one spot lower than Illinois despite having no big resume building wins. I really think ESPN just wants to see this as part of the equation on selection sunday just as they have with Bracketology.

All about the site hits. Once they get the system aligned properly, they'll be able to insert themselves into the conversation.

What Dayton said really put it into perspective. They're treating games in the way a teacher would hand out and A/B/C/D/F grades in a class, when in reality games are all Pass/Fail classes. Does Duke's championship mean less because they only won by two? No.

[ILLINISTU]

Originally Posted by IlliniRyan13

Total garbage. Lunardi has also been horrific as a bracketologist past few years. So the worse he seeds us, the better I feel.

+1000 He usually gets most of the teams right, but that seeding is rediculous. Butler and KState below New Mexico...? I didn't bother to read it all, made me mad.