Bubble Rap 2013

[Kedsy]

I suppose there is some degree of self-reinforcement here – maybe one reason that Savannah St., Florida and Texas Southern’s non-con SOS rankings are so high is that they played MTSU, which won 28 games? And then that feeds back into making MTSU’s SOS look better?

I'm not 100% certain, but I don't think that's exactly how the RPI SOS works. I mean yes, anyone who played MTSU gets a boost in its non-con SOS (your first question), but I don't think it feeds back. Because as I understand it, SOS is the aggregate winning percentage of your opponents. The only way it could feed back is if opponents' SOS contributed to your SOS. And while opponents' SOS is part of your RPI, I don't believe it's part of your SOS. I know that sounds confusing, but if you parse it out I'm pretty sure the answer to your second question is no it doesn't.

Having said all that, in my opinion the fact that MTSU's non-conference SOS is 11th ought to be good enough to get a team with a top 30 RPI into the Dance.

[JasonEvans]

Having said all that, in my opinion the fact that MTSU's non-conference SOS is 11th ought to be good enough to get a team with a top 30 RPI into the Dance.

We all know it is possible to "game" the RPI and most of the SOS measurements. You do it by playing a lot of decent but not great teams in your non-conference slate and making sure to avoid really horrible ones. To me, the real measure of a SOS is that when I look down at your schedule, I see a fair number of quality teams on there who are solid tournament teams.

I am not certain, because there is a lot of math to make it all work, but I believe the following scenarios would be fairly similar schedules from a SOS metric standpoint.

Team A plays 10 games- 3 against teams in the Top 25, 2 against teams in the Top 26-50, and 5 against teams outside the top 250.
Team B plays 10 games - 2 against teams in the 50-100 range, 6 against teams in the 100-175 range, and 2 against teams in the 175-200 range.

If you ask me, there is no comparison in those schedules. Team A really faced some top flight competition! Team B just did not.

Look, for a team in the top 50 or 75, beating a team that is ranked 175 or a team that is ranked 275 is about the same thing. Neither stands much of a chance of toppling you. But, fill your schedule with those teams that are 175 and it can look like you played a tough schedule compared to a team that has a few super-gimmies against teams ranked around #300.

You guys keep pointing to MTSU's 11th best SOS and saying it is good enough, but I keep on looking at it and saying, "but they didn't beat any good teams!!" Their only top 100 win is Ole Miss!!!!! THEY HAVE ONE VICTORY OVER TEAMS IN THE TOP 100!!!! How can you say they deserve a seat at the Dance when they haven't beaten anyone else who will be there!!?!?! What's more, the only 2 times they played NCAA Tournament teams, Florida and Belmont, they got beaten by 15+ points.

I truly think some of ya'll are getting too caught up in "metrics" and looking at their SOS without looking past it to see the actual teams they played... and the teams they beat.

The NCAA tournament is pretty much nothing but games against teams in the top 60 or so in the RPI. Sure, there may be a couple auto-qualifiers outside that range, but unless you are a #1 seed who has earned the right for an easy opening round game, you are getting a top 50 opponent. MTSU versus the Top 50 is 0-2. Every single other Bubble team has a top 50 win... most of them have several of them. Every single other bubble team has multiple top 100 wins. Not MTSU. Their lack of quality wins sticks out like a sore thumb. If you can't beat the best, don't tell me you deserve to play with the best.

-Jason "I am enjoying this-- especially because I am convinced the selection committee will agree with me and leave MTSU at home" Evans

[Wander]

We all know it is possible to "game" the RPI and most of the SOS measurements.

Reminds me of an xkcd comic: http://xkcd.com/940/. Hover your mouse over the comic and the following hidden text comes up: "I felt so clever when I found a way to game the Fitocracy system by incorporating a set of easy but high-scoring activities into my regular schedule. Took me a bit to realize I'd been tricked into setting up a daily exercise routine."

At any rate, would you agree that it's harder to game Pomeroy than the RPI? Because they're 31st in that, ahead of many other bubble teams.

[JasonEvans]

At any rate, would you agree that it's harder to game Pomeroy than the RPI? Because they're 31st in that, ahead of many other bubble teams.

Yup, harder but still not impossible. I just keep on coming back to the fact that they beat no one who will get an at-large bid. Heck, unless Ole Miss plays well in the SEC tourney, The Rebs won't even on the bubble. So, MTSU would have zero victories over bubble teams or better. You gotta beat someone to make the dance. They didn't.

-Jason "I sound like a broken record, don't I?" Evans

[Kedsy]

We all know it is possible to "game" the RPI and most of the SOS measurements. You do it by playing a lot of decent but not great teams in your non-conference slate and making sure to avoid really horrible ones. To me, the real measure of a SOS is that when I look down at your schedule, I see a fair number of quality teams on there who are solid tournament teams.

Yes, we've all had this conversation before, about whether a schedule with some top teams and some bottom teams is stronger or weaker than a schedule with a bunch of middle teams. Obviously there's no clear right or wrong answer to that question, although of course everyone's entitled to their opinion in that regard.

What bothers me in this debate is the idea that we can use the numbers if we agree with them, but when our eyes tell us the numbers are "wrong" we should disregard them. The committee adopted computer rankings because the eye test was a poor way to fill the field. Either we should use the numbers we have or if we don't think they're right we should get new numbers. The RPI is the system the committee has said for years is the yardstick (although fortunately they seem to be slowly moving away from that), so if MTSU has a strong RPI and RPI SOS it's not right for the committee to say MTSU didn't play a good schedule and is thus undeserving. Especially since it's not just the RPI but every other computer system as well that suggest MTSU is good enough to make the field.

You guys keep pointing to MTSU's 11th best SOS and saying it is good enough, but I keep on looking at it and saying, "but they didn't beat any good teams!!" Their only top 100 win is Ole Miss!!!!! THEY HAVE ONE VICTORY OVER TEAMS IN THE TOP 100!!!! How can you say they deserve a seat at the Dance when they haven't beaten anyone else who will be there!!?!?! What's more, the only 2 times they played NCAA Tournament teams, Florida and Belmont, they got beaten by 15+ points.

I truly think some of ya'll are getting too caught up in "metrics" and looking at their SOS without looking past it to see the actual teams they played... and the teams they beat.

The NCAA tournament is pretty much nothing but games against teams in the top 60 or so in the RPI. Sure, there may be a couple auto-qualifiers outside that range, but unless you are a #1 seed who has earned the right for an easy opening round game, you are getting a top 50 opponent. MTSU versus the Top 50 is 0-2. Every single other Bubble team has a top 50 win... most of them have several of them. Every single other bubble team has multiple top 100 wins. Not MTSU. Their lack of quality wins sticks out like a sore thumb. If you can't beat the best, don't tell me you deserve to play with the best.

Well, first of all, Akron is a probable NCAA team and MTSU played them in Akron and took them to OT. Mississippi also has a decent chance to be an NCAA team, unless of course MTSU beats them out. Moreover, Mississippi is currently #56 in the RPI (#45 in Pomeroy), so it's silly to say every other bubble team has a top 50 win and MTSU doesn't, unless you think there's some huge difference between #49 and #56. And of course the same for the top 100 (one of the teams MTSU beat is currently ranked #107, for example).

Also, you seem fine using the RPI to determine things like "top 50 wins," but that's inconsistent. Schedule strength, calculated in a way you think is incorrect, makes up 75% of the RPI. So unless you've analyzed the schedules of all those vanquished top 50 teams (to make sure they pass the eye test and are "really" top 50), the bubble teams may not have beaten "real" top 50 teams at all. Where does that leave us?

Finally, you appear to want making the NCAA tournament to be almost solely about big wins. TCU has a big win against Kansas, does that make them more worthy than MTSU? Of course not. "Deserving" to play in the NCAA tournament is about who had the better season, not who won the biggest single game (except I guess when it comes to conference tournaments, where it becomes about who won the conference final). MTSU's entire season put them in the top 30 of both the RPI and Pomeroy (OK, currently #31 in Pomeroy), despite the handicap of playing in a conference that drags down its SOS (and thus its overall rating). It doesn't seem right to add additional requirements because you don't believe the ratings.

[hurleyfor3]

At any rate, would you agree that it's harder to game Pomeroy than the RPI?

I think Billy Donovan has figured out how to game Pomeroy. They have a lot of wins with scores like 66-40 and 75-36. You can beat South Carolina that way but I doubt you can do it against most of the Top 25, except perhaps Wisconsin.

[loran16]

I think Billy Donovan has figured out how to game Pomeroy. They have a lot of wins with scores like 66-40 and 75-36. You can beat South Carolina that way but I doubt you can do it against most of the Top 25, except perhaps Wisconsin.

That's not gaming Pomeroy. They've been doing that to their conference - they didn't schedule those games. if other teams could beat their conference by 20 points per game, they'd do it to.

[robed deity]

That's not gaming Pomeroy. They've been doing that to their conference - they didn't schedule those games. if other teams could beat their conference by 20 points per game, they'd do it to.

While that's true, it is evident that slower paced teams have a slightly inflated pomeroy rating. Florida plays at a slow pace and blows mediocre teams out, which accounts for the high pp statistics. Florida is a really good team, but I'm not sure they are really the best team in the country, as kenpom has had them most of the year. Similarly, I'm not sure Denver and UVA (both slow paced teams) are as good as their ratings suggest.

[hurleyfor3]

That's not gaming Pomeroy. They've been doing that to their conference - they didn't schedule those games. if other teams could beat their conference by 20 points per game, they'd do it to.

OK, maybe Billy isn't literally trying to game Pomeroy, but I maintain they are not as impressive as their dork poll numbers suggest. EVERY game they've played where the margin of victory was less than 12 points has been a loss. I'm rather amazed, as I don't even think 1991 unlv accomplished that. It doesn't imply a deep ncaa tournament run.

[Reilly]

... Mississippi is currently #56 in the RPI (#45 in Pomeroy), so it's silly to say every other bubble team has a top 50 win and MTSU doesn't... .

Ole Miss is #33 in the www.sports-reference.com "SRS":

http://www.sports-reference.com/cbb/...standings.html

MTSU is #48 per s-r.

[Wander]

Yup, harder but still not impossible. I just keep on coming back to the fact that they beat no one who will get an at-large bid. Heck, unless Ole Miss plays well in the SEC tourney, The Rebs won't even on the bubble. So, MTSU would have zero victories over bubble teams or better. You gotta beat someone to make the dance. They didn't.

Hey, for the record, I strongly agree with you that playing a tough schedule means absolutely nothing if you don't do well against that schedule. I think I just differ in that I put more value in consistently beating the decent-to-mediocre teams than you do.

As always, we can't consider this in a vacuum. We have to look at what other teams are doing. And the big catch in doing that is that asking "Who did you beat?" is only half the equation; to me, "who did you lose to?" is equally as important. Virginia has much better wins than MTSU does; they also have 2 losses that are about as bad or worse that MTSU's worst loss, and 7 or so losses that are about as bad or worse than MTSU's second worst loss. Like Kedsy said, people can reasonably differ over which is better - having both good wins and bad losses, or having lots of mediocre wins - but I don't think one answer is super obvious over another. To be honest I'm undecided on MTSU.

[sporthenry]

Finally, you appear to want making the NCAA tournament to be almost solely about big wins. TCU has a big win against Kansas, does that make them more worthy than MTSU? Of course not. "Deserving" to play in the NCAA tournament is about who had the better season, not who won the biggest single game (except I guess when it comes to conference tournaments, where it becomes about who won the conference final). MTSU's entire season put them in the top 30 of both the RPI and Pomeroy (OK, currently #31 in Pomeroy), despite the handicap of playing in a conference that drags down its SOS (and thus its overall rating). It doesn't seem right to add additional requirements because you don't believe the ratings.

Well this is going to be the crux of the bubble. Do you put a team in who was very consistent against mediocre opponents but hasn't beaten an at-large team? Or do you put a wholly inconsistent team who has multiple wins against top 25 teams but also losses versus some terrible teams.

Personally, I prefer a team who can beat the big boys. I don't see how you put MTSU over a team like ISU. Perhaps it is b/ MTSU hasn't had as many chances but they got blown away by UF and then have 3 other games versus teams who are probably on the wrong side of the bubble. So they can beat or compete with other 12 seeds, ISU can beat and compete with 2,3,4 seeds.

[OldPhiKap]

Well this is going to be the crux of the bubble. Do you put a team in who was very consistent against mediocre opponents but hasn't beaten an at-large team? Or do you put a wholly inconsistent team who has multiple wins against top 25 teams but also losses versus some terrible teams.

Personally, I prefer a team who can beat the big boys. I don't see how you put MTSU over a team like ISU. Perhaps it is b/ MTSU hasn't had as many chances but they got blown away by UF and then have 3 other games versus teams who are probably on the wrong side of the bubble. So they can beat or compete with other 12 seeds, ISU can beat and compete with 2,3,4 seeds.

I agree that, for at-large berths, you want folks who can go out and compete with the big guys.

Consistency is great as well, which is what the tournament championship should typically produce. So they know what they need to do to get in.

[Kedsy]

Well this is going to be the crux of the bubble. Do you put a team in who was very consistent against mediocre opponents but hasn't beaten an at-large team? Or do you put a wholly inconsistent team who has multiple wins against top 25 teams but also losses versus some terrible teams.

Personally, I prefer a team who can beat the big boys. I don't see how you put MTSU over a team like ISU. Perhaps it is b/ MTSU hasn't had as many chances but they got blown away by UF and then have 3 other games versus teams who are probably on the wrong side of the bubble. So they can beat or compete with other 12 seeds, ISU can beat and compete with 2,3,4 seeds.

By ISU, I assume you mean Iowa State? I just looked at Iowa State's schedule, which includes two (2) out-of-conference games against top 50 RPI teams (both losses for Iowa State). It also includes eight conference matchups against top 50 RPI teams, and Iowa State won three of the eight, winning home games against Oklahoma, Oklahoma State, and Kansas State (but losing road games against all three teams) plus two losses to Kansas.

First of all, I don't see a 2 or 3 seed there that Iowa State competes with. Although K State, which is Iowa State's only top 25 win, and Oklahoma State are both #4s in Lunardi's latest bracket. Second, all of Iowa State's top 50 wins are at home and all are against conference opponents (with which presumably Iowa State has some familiarity). MTSU can't get that sort of game from within its conference. And the only top 50ish team MTSU apparently was able to convince to visit them was Ole Miss, and MTSU won.

Finally, you say you'd rather have a team with some good wins and some bad losses, but how far does that go? If Iowa State had four more medium/bad losses, giving them a 17-14 overall record (8-10 Big 12) would you still put them in over MTSU? They'd still have exactly the same number of wins against the big boys. Assuming your answer is "no," then it's just a matter of degree -- you're still looking at the whole season and you're saying all other things being equal you'd pick the team with better wins, right? It's like a tie-breaker. So now it's just a matter of how you tell if all other things are equal. Which supposedly is where the computer rankings come in. If we compare Iowa State and MTSU in the RPI, MTSU is ranked #29 while Iowa State is #47 (it's closer in Pomeroy: #31 for MTSU and #36 for Iowa State). To me, that doesn't look so equal that we'd have to pull out the tie-breaker, but obviously reasonable minds can differ about something like that.

Ultimately, I don't really feel strongly about MTSU's at-large candidacy. What I do feel strongly about is if you choose a computer system to avoid being overly subjective, you probably shouldn't ignore the computer when it doesn't agree with your eye test.

[WakeDevil]

This Jason says another Jason doesn't know what he's talking about. It's not easy to compare an MT to a VA, so he developed a system he explains in the column.

http://www.thebiglead.com/index.php/...-bubble-teams/

When we try to account for Simpson’s paradox, understanding that Middle Tennessee and St. Mary’s played fewer top teams and a higher percentage of teams below 100, we see that both teams still outperformed their bubble comrades at every level from teams ranked 51-100, to those ranked 101-200, to the 201+ category. Neither of them played a team in the 26-50 category in RPI, so we have no comparison to draw there.

For the top group, St. Mary’s has one win in four opportunities, roughly in line with the bubble teams, while Middle Tennessee went 0-2. A-ha! See Middle Tennessee is not worthy!

Not so fast. Middle Tennessee lost to Florida in a road/neutral game in Tampa populated by a pro-Gator crowd, and at Belmont. No one traveled into their arena, except for Ole Miss, one of those bubble teams, and Middle Tennessee won that one. So if we want to really compare Middle Tennessee’s record against top 25 teams, it needs to be truly comparable, by accounting that their games were on the road.

Those 14 bubble teams that the committee will be debating went 1-33 against the RPI top 25 in road/neutral games. No, seriously, they have. That’s dreadful. Middle Tennessee going 0-2 doesn’t seem like a strike that knocks them down now, nor does St. Mary’s losing at Gonzaga and again in Vegas.

[sporthenry]

When we try to account for Simpson’s paradox, understanding that Middle Tennessee and St. Mary’s played fewer top teams and a higher percentage of teams below 100, we see that both teams still outperformed their bubble comrades at every level from teams ranked 51-100, to those ranked 101-200, to the 201+ category. Neither of them played a team in the 26-50 category in RPI, so we have no comparison to draw there.

My whole problem with this though, is that it also fails to take into account the human factor. On paper, the UVA and UMD losses are similar to if Gonzaga lost to BYU or Saint Mary's depending upon whether you want to use Kenpom, RPI, Sagarin, etc. But this fails to take into account such things as Duke was just playing their 3rd game in 6 days against ACC talent including beating their arch rival. Or that at UVA was right before their big game versus Miami.

Gonzaga got to play some horrid teams who don't really pose a threat to them in teams like Portland or Pepperdine. So their trap games were a Portland or a Loyola Marymount. That would be like Duke getting Elon or Cornell before or after some of their biggest games. For the record, Duke beat Elon 76-54 and Cornell 88-47. And their "big" game would have been Duke's 5th or 6th biggest game in their conference.

So I'm willing to take a team that might underperform in the 50-100 or 100-200 range b/c of the strength of their schedule. His WAMAL is an interesting concept b/c he at least seems to recognize that the difference between 200 and 300 isn't huge while the difference between 5 and 50 is huge, at least for our purposes but I'll have to look into it more.

But I really can't get behind something that is willing to put MTSU ahead of a team like Nova. I'd love to see MTSU go from beating Marquette, a top 25 team and then have to play Seton Hall. According to kenpom, Seton Hall is better than any other team in the Sun Belt and the best team MTSU would have faced since December.

Now sure, a big win doesn't put TCU in the tourney just the same as consistency in conference doesn't put Norfolk State in there, neither side is really arguing that so bringing them up is more of a straw man. But when all else is about equal, I think the big win should put you over. And a win versus Syracuse and a loss to Seton Hall is better than a loss to Syracuse and a win versus Seton Hall.

[sporthenry]

First of all, I don't see a 2 or 3 seed there that Iowa State competes with. Although K State, which is Iowa State's only top 25 win, and Oklahoma State are both #4s in Lunardi's latest bracket. Second, all of Iowa State's top 50 wins are at home and all are against conference opponents (with which presumably Iowa State has some familiarity). MTSU can't get that sort of game from within its conference. And the only top 50ish team MTSU apparently was able to convince to visit them was Ole Miss, and MTSU won.

Finally, you say you'd rather have a team with some good wins and some bad losses, but how far does that go? If Iowa State had four more medium/bad losses, giving them a 17-14 overall record (8-10 Big 12) would you still put them in over MTSU? They'd still have exactly the same number of wins against the big boys. Assuming your answer is "no," then it's just a matter of degree -- you're still looking at the whole season and you're saying all other things being equal you'd pick the team with better wins, right? It's like a tie-breaker. So now it's just a matter of how you tell if all other things are equal. Which supposedly is where the computer rankings come in. If we compare Iowa State and MTSU in the RPI, MTSU is ranked #29 while Iowa State is #47 (it's closer in Pomeroy: #31 for MTSU and #36 for Iowa State). To me, that doesn't look so equal that we'd have to pull out the tie-breaker, but obviously reasonable minds can differ about something like that.

As for a 2 seed that ISU has competed with, how about Kansas? They played them to OT both home and away. And K-State/OSU could conceivably jump up to the 3 line mainly b/c the 3/4 seed lines seem very fluid.

As for the computer systems, I don't think anyone is saying to follow them blindly. If anything, that has been a huge problem with the committee and people have wanted to use the eye test to supplement the computer candidacy. I know I pull up computer rankings quite a bit on here but I just to use them as a starting point and more so for groupings such as previously when I tried to identify similar teams to Maryland or UVA. I'm not really comfortable saying that this team at 30 must be better than the team at 35 especially when we have acknowledged that his system isn't flawless with the Wisconsin problem. And that doesn't even take into account people using the system without even knowing how the numbers are derived (not saying you and I was guilty of this until recently). Oh and for the record, Sagarin has ISU at 36 and Middle Tennessee at 55 and BPI has them at 34 with MTSU at 46.

But I agree that when looking at the teams, it comes down to how much emphasis you put on the consistency of beating the teams you are supposed to compared with getting top wins. The article linked earlier about WAMAL seemed like a good starting point b/c it accounts for both consistency as well as strength of opponents in the sense that it doesn't throw in all top 25 wins into the same basket but that seemed to heavily favor mid-majors beating teams they were supposed to and didn't really value big wins as much as I would and didn't really take into account that playing 3 average bubble teams is much tougher than playing 1 bubble team and 2 terrible teams.

But this is the beauty of college basketball, we are arguing about MTSU and ISU who will both be home by the first weekend so in the big scheme of things, this really doesn't matter much.

[Reilly]

... But this is the beauty of college basketball, we are arguing about MTSU and ISU who will both be home by the first weekend so in the big scheme of things, this really doesn't matter much.

And a couple years ago, we were arguing about VCU who also would be home by the first weekend so it didn't really matter much except they didn't go home and so it did matter ... which makes it all the more beautiful.

[pfrduke]

I don't think Providence was actually on the bubble, but we can definitely put a stake through their at large campaign after a 17-point loss to Cincinnati to open the Big East tourney.

[tommy]

Interesting discussion about the merits of Middle Tennessee State and the philosophy and priorities of the committee. Seems like it's going to come down to whether the committee is inclined to give credit to a mid or low major for scheduling games against high level teams, or whether it wants to see them actually beat some of those teams. Credit for trying, or for succeeding? Kind of like the coach, when asked about whether his team has any three point shooters, responding "we've got plenty of three point shooters. What we don't have is any three point makers!"

It may or may not be instructive to look at what the committee has done in the recent past. For instance, last year, look at a team like Iona. The Gaels were the best team in the regular season in the MAAC last year, going 24-6 in the regular season and 15-3 in the league. But they were upset in the semifiinals of the conference tournament by Fairfield, who then lost in the finals to Loyola (Md.) who of course got the automatic bid. But what about Iona?

Going 1-1 in the MAAC tournament left them at 25-7. Their RPI was 42 -- pretty good. SOS though was only 163. Why? They only played two games against the top 50 (and they lost them both). They were 5-1 against the RPI 51-100. Nonconference SOS was 44.

What about the eye test vis-a-vis the schedule? They did blow out Maryland, but that was in November, and Maryland stunk last year. The only other game against a BCS squad was the season opener in November against Purdue, which the Gaels lost by a point. The Boilers turned out to be pretty good, winning a NCAA Tournament game and ending up at #44 in the RPI. Iona beat Richmond, but they didn't make the postseason, and they beat Nevada, who was the best team in their league but got upset in their conference tournament and had to settle for the NIT. Iona also beat St. Joseph's, an NIT team. Not much else. Other losses included Marshall, Hofstra, Manhattan, Siena, and Loyola (Md.).

Of course the admissibility of any team depends on the teams in that particular year that they're being compared to. I would think that Jason would have said "no way" to Iona. They didn't beat anybody who mattered, and didn't demonstrate the ability to beat teams the quality of which they'd be playing in the tournament.

The committee extended the Gaels an at-large bid, seeding them 14th -- as low as any at-large team has ever been seeded-- and they lost to BYU in a first round play-in game. This team didn't appear to schedule particularly difficult teams out of conference, and the one tournament team they played -- Purdue -- they lost to. But they got in anyway.