Ranking D3 BBall Conferences

Mr. Ypsi · October 12, 2009, 03:10:06 PM

It would at least indicate strength at the top of the conference. But, to me at least, if you are ranking conferences, you've got to consider the entire conference.

Another problem with overly relying on tourney results is that with a one-loss-you're-out format, successful teams have to be not only good, but also fortunate. In IWU's only title year (1997) it took a falling-down final-seconds shot to survive Rose Hulman in the second round (a team we had beaten by 25 earlier in the season). Greg has mentioned that at least 3 of NPU's 5 national title teams had at least one game that they were 'fortunate' to win. (Personally, I'm not sure IWU's title team was even among the top FIVE IWU teams all-time - but they are the ones who prevailed.)

Gregory Sager · October 13, 2009, 06:11:05 PM

Quote from: Mr. Ypsi on October 12, 2009, 03:10:06 PMAnother problem with overly relying on tourney results is that with a one-loss-you're-out format, successful teams have to be not only good, but also fortunate. In IWU's only title year (1997) it took a falling-down final-seconds shot to survive Rose Hulman in the second round (a team we had beaten by 25 earlier in the season). Greg has mentioned that at least 3 of NPU's 5 national title teams had at least one game that they were 'fortunate' to win. (Personally, I'm not sure IWU's title team was even among the top FIVE IWU teams all-time - but they are the ones who prevailed.)

Last season's Wash U outfit was a classic example of that. The Bears won their first three tourney games by two, three, and three points -- and in their first-round game against Lawrence they actually trailed going into the final twenty seconds.

In order to compare the CCIW and the UAA on a top-to-bottom basis over the past four years based upon head-to-head results, I jotted down the results of all of the games between the two leagues -- which, of course, mostly involve Chicago and Wash U on the UAA side -- and the standings of the two leagues for each season. Then I went through the results and assigned a value of +1 to either league if that league's team beat the other league's standings counterpart head-to-head (e.g., UAA #1 Wash U beating CCIW #1 Wheaton last season), +1.5 if that league's team beat the team that finished one place farther up in the other league, +2 if that league's team beat the team that finished two places farther up in the other league, etc. I didn't assign any points for wins over a team that finished farther down in the other league, since, if the two leagues are presumed to be equal, such a victory would be an expected outcome.

Here's what I came up with:

2008-09
Elmhurst (CCIW #2) > Washington (UAA #1)
North Park (CCIW #8) > Case Western Reserve (UAA #6)
Illinois Wesleyan (CCIW #7) > Chicago (UAA #5)
Washington (UAA #1 ) > Wheaton (CCIW #1)
total: CCIW 5.5, UAA 1

2007-08
Wheaton (CCIW #2) > Chicago (UAA #1)
Washington (UAA #2) > Illinois Wesleyan (CCIW #2)
Washington (UAA #2) > Augustana (CCIW #1)
total: UAA 2.5, CCIW 1.5

2006-07
Augustana (CCIW #1) > Washington (UAA #1)
total: CCIW 1, UAA 0

2005-06
Wheaton (CCIW #5) > Chicago (UAA #3)
Illinois Wesleyan (CCIW #2) > Washington (UAA #2)
total: CCIW 3, UAA 0

four-year total: CCIW 11, UAA 3.5

That's not definitive, but it's a good snapshot of how the two leagues have measured up to each other on a slot-by-slot basis.

Hugenerd · October 13, 2009, 08:10:00 PM

What a wonderful extremely arbitrary system, why not just make it an exponential dependence on difference, that would make the CCIW look even better. Also only one game respresents teams other than WashU and Chicago (great sample size) and you essentially discount all of the UAAs wins because you dont take into account essentially all of WashU's wins because they have been #1 or #2 in the UAA for the last 4 years and Chicago has been #1 two of those years and 3rd another year. So a win by WashU on the road @ #3 Augie doesnt count for anything? I also noticed you took the liberty of rounding the rankings up in the case of a tie. This may not seem like a big deal, but WashU was tied for 2nd in 2008 (so effectively 2.5) and tied for first in 2007, so they were not solely 1st and 2nd. Therefore, since every game but one was played by WashU and Chicago and those teams have accounted for 4 #1s (they tied in 2007), 2 #2s, and 1 #3, nearly all of their wins are discounted and they are punished very heavily for an upset.

Just to give you an example of what I mean. If WashU played every team in the CCIW each year from 2005 to 2008 and they were number #1 in the UAA each year, there would be only 1 game each year where they could gain points, whereas the CCIW would have 8 games each year to gain points. If the CCIW lost 7 of those games and won only 1, lets say an upset by the #3 team, WashU would still be -1 overall per year(1 pt for beating CCIW #1 team, -2 for losing to #3 team), even though they went 7-1 each year, and -4 for the total of the 4 years. I am sorry, but this may be the worst comparison/analysis/invention-of-a-system-to-skew-an-argument-in-your-favor I have ever seen. If anyone would like to rank the CCIW over the UAA because of this argument, than that is your right, but I am sure that I can come up with an arbitrary analysis that shows the UAA equally superior to the CCIW (for example, tourney wins vs. CCIW).

Mr. Ypsi · October 13, 2009, 09:45:14 PM

While Greg's system had plenty of flaws (as he already acknowledged), it did use the available data. You complain about his 'small sample size', yet propose to substitute tourney wins (which would reduce the UAA sample recently to ONE, and losing to WashU hardly makes the CCIW unique)!

(Besides, if you wanna go tourney I'll pull out the 2001 card: when CCIW #3 IWU beat Chicago en route to the Final Four, Chicago was not just UAA #1, they were ranked #1 in the country.

)

Hugenerd · October 13, 2009, 10:22:48 PM

Quote from: Mr. Ypsi on October 13, 2009, 09:45:14 PM
While Greg's system had plenty of flaws (as he already acknowledged), it did use the available data. You complain about his 'small sample size', yet propose to substitute tourney wins (which would reduce the UAA sample recently to ONE, and losing to WashU hardly makes the CCIW unique)!

(Besides, if you wanna go tourney I'll pull out the 2001 card: when CCIW #3 IWU beat Chicago en route to the Final Four, Chicago was not just UAA #1, they were ranked #1 in the country. )

Actually, two in the past two years, they won @ Augie 2 years ago and @ Wheaton last year. If I use the following formula to calculate an arbitrary number of points: 1 point for win, +5 for win on the road, + (difference in national ranking entering tourney)^2 if winner is ranked lower than opponent.

Here is what we have

WashU
2008: 1+5+(11-6)^2= 31 (win @ Augie)
2009: 1+5+0 = 6 (win @ Wheaton)

Under my system: the UAA clearly has the advantage over the CCIW 37-0

As I said almost a dozen times now, you can make any argument you want with the small sample size. Therefore, especially because of the regional issues raised in d3, no one is going to make any headway in this argument.

Mr. Ypsi · October 13, 2009, 10:36:30 PM

I meant a sample size of one team (WashU), not one game. Sorry for the lack of clarity.

Of course this is unresolvable - like most sports arguments! That's why they're fun!

Hugenerd · October 13, 2009, 11:02:19 PM

Quote from: Mr. Ypsi on October 13, 2009, 10:36:30 PM
I meant a sample size of one team (WashU), not one game. Sorry for the lack of clarity.

Of course this is unresolvable - like most sports arguments! That's why they're fun!

That is quite alright, this entire argument lacks clarity.

I guess my main disagreement with the general "top-to-bottom" argument is that no conference ever has 0 weak teams. Nobody is making the argument that the Big East isnt the top D1 basketball conference because DePaul, Rutgers and South Florida have been absolutely horrible (combined 30% winning percentage last year, South Florida lost badly to Oral Roberts, Wright State, and Niagara, Rutgers lost to Lehigh, Binghamton, etc.). Also, as is the case with poor teams, they are usually inconsistent. So if Case lost to North Park on any given night, that is not surprising to me (WashU beat North Park by 24, but didnt beat Case by more than 10 points in 2 meetings). It is one game, which unfortunately is a large basis of the "top-to-bottom" argument. The same Case team was also beating WashU for 38 minutes last year (Emory, the worst team in the conference was also within a basket of beating WashU). And that is not to say that North Park isnt a better team than Case Western, they may be. All I am saying is that just because the bottom 3 teams in the CCIW are better than the bottom 3 in the UAA, that does not make it impossible for the UAA to be a better conference overall than the CCIW. That is why I think that, although top-to-bottom strength is one factor in overall conference strength, another essential factor is the strength of the other teams in the conference, especially those at the top.

Gregory Sager · October 14, 2009, 06:52:47 PM

Quote from: hugenerd on October 13, 2009, 08:10:00 PM
What a wonderful extremely arbitrary system, why not just make it an exponential dependence on difference, that would make the CCIW look even better. Also only one game respresents teams other than WashU and Chicago (great sample size) and you essentially discount all of the UAAs wins because you dont take into account essentially all of WashU's wins because they have been #1 or #2 in the UAA for the last 4 years and Chicago has been #1 two of those years and 3rd another year. So a win by WashU on the road @ #3 Augie doesnt count for anything?

As I said, if you start from the neutral assumption that the two leagues are presumed to be equal, a #1 team should beat a #3 team. Yes, road factors can cancel that out -- in the CCIW, even the #1 team often can't beat the #3 team in the #3 team's gym -- but you can't quantify them. More to the point, your complaint that the experiment is skewed against the UAA because Wash U and Chicago are perennially at the top of that circuit really doesn't hold. Of the games listed, the UAA only has a #1 represented one more time than does the CCIW, 4 to 3.

Quote from: hugenerd on October 13, 2009, 08:10:00 PMI also noticed you took the liberty of rounding the rankings up in the case of a tie. This may not seem like a big deal, but WashU was tied for 2nd in 2008 (so effectively 2.5) and tied for first in 2007, so they were not solely 1st and 2nd.

That argument cuts both ways. Wheaton and Wesleyan tied for second in the CCIW in 2007-08, and Wesleyan was one of three teams that tied for second in 2005-06.

Quote from: hugenerd on October 13, 2009, 10:22:48 PMAs I said almost a dozen times now, you can make any argument you want with the small sample size. Therefore, especially because of the regional issues raised in d3, no one is going to make any headway in this argument.

Yes, the regional issues that constitute most of the UAA's strengths in this argument.

Quote from: hugenerd on October 13, 2009, 11:02:19 PM
Quote from: Mr. Ypsi on October 13, 2009, 10:36:30 PM
I meant a sample size of one team (WashU), not one game. Sorry for the lack of clarity.

Of course this is unresolvable - like most sports arguments! That's why they're fun!

That is quite alright, this entire argument lacks clarity.

I guess my main disagreement with the general "top-to-bottom" argument is that no conference ever has 0 weak teams. Nobody is making the argument that the Big East isnt the top D1 basketball conference because DePaul, Rutgers and South Florida have been absolutely horrible (combined 30% winning percentage last year, South Florida lost badly to Oral Roberts, Wright State, and Niagara, Rutgers lost to Lehigh, Binghamton, etc.).

Stick to the apples-and-apples analogies, please. There are plenty of D3 leagues to choose from; let's leave D1 out of this.

The ideal examples have already been introduced to this discussion by Chuck: The MIAA and the NCAC. Nobody disputes the fact that, year in and year out, the MIAA has two of the strongest programs in the land in Hope and Calvin. Yet, year in and year out, the MIAA posts a losing cumulative non-con record. The NCAC is more of the same; Wooster and Wittenberg are consistently among the top teams in D3, and yet the NCAC posts one bad non-con season after another.

Your contention that "no conference ever has zero weak teams" is demonstrably false. Look at the CCIW; North Park went winless in league play last season, and yet the Vikings went 8-3 in non-con play, with one of the three losses coming at the hands of a D1 team. In each of the past three seasons the CCIW's bottom-dweller has posted a winning record in non-con play. The WIAC has done it for four seasons in a row.

Quote from: hugenerd on October 13, 2009, 11:02:19 PMAlso, as is the case with poor teams, they are usually inconsistent. So if Case lost to North Park on any given night, that is not surprising to me (WashU beat North Park by 24, but didnt beat Case by more than 10 points in 2 meetings). It is one game, which unfortunately is a large basis of the "top-to-bottom" argument.

It's actually two games (CCIW #7 IWU beat UAA #5 Chicago), and in both cases the CCIW team beat a UAA team that finished two notches higher. But I'm more than willing to go by the much larger sample size afforded by cumulative non-con record, with the understanding that three of the UAA's eight schools -- Brandeis, Rochester, and NYU -- play in regions that are vastly inferior to the Midwest Region, and that this substantial number of non-con games with East and Northeast teams distorts the UAA's final tally for comparative purposes. In that respect, the UAA's overall advantage over the past four years, which is 287-106 (.730) to the CCIW's 268-104 (.720), seems very, very underwhelming.

Quote from: hugenerd on October 13, 2009, 11:02:19 PMThe same Case team was also beating WashU for 38 minutes last year (Emory, the worst team in the conference was also within a basket of beating WashU). And that is not to say that North Park isnt a better team than Case Western, they may be.

The Vikings led for the last 29 minutes of the game, the last nine minutes in double digits, and they did it with two starters sitting on the bench in street clothes due to injury. NPU was clearly a better team than CWRU.

Conference games tend to take on different patterns than non-con games, especially on the road. The competition is typically much fiercer, due to both higher stakes and greater familiarity, and blowouts tend to be less common. It's not at all unusual in any league to find a bottom-dweller giving a contender a tough game; that's what conference play is all about. What's ultimately important is that the better team usually wins, so the ability of a bottom-dweller to hang with a top team for 38 minutes before succumbing is really a distinction without a difference.

Quote from: hugenerd on October 13, 2009, 11:02:19 PMAll I am saying is that just because the bottom 3 teams in the CCIW are better than the bottom 3 in the UAA, that does not make it impossible for the UAA to be a better conference overall than the CCIW.

No, but it makes it very difficult, because you've conceded three out of the eight spots. That only leaves you five left, and the UAA's gotta be better at four of those five just to break even.

Quote from: hugenerd on October 13, 2009, 11:02:19 PMThat is why I think that, although top-to-bottom strength is one factor in overall conference strength, another essential factor is the strength of the other teams in the conference, especially those at the top.

What "other teams"? Top-to-bottom strength means just that: Every team in the conference, which in the case of both the CCIW and UAA means all eight of them.

Your contention that strength at the top taken in isolation is an "essential factor" just doesn't make sense. It disregards the fact that there are eight teams in the league, not one or two. And, as both Chuck and I have pointed out, it leaves you holding the bag when the MIAA and the NCAC are brought up as counterexamples.

Mr. Ypsi · October 14, 2009, 08:15:57 PM

Let me introduce a new angle (admitting right away that there is probably no conclusive game-results to test the proposition): granting that every conference can be forgiven ONE deadbeat (though Greg already showed that the last place teams in the CCIW and WIAC still consistently post winning non-con records), how strong are your 'near-bottom' teams?

While (as I admitted) there is probably no real data to 'prove' it, I doubt you'd find many (if any) national observers claiming that any other conference's #6 and #7 teams (or their equivalent in larger or smaller conferences) would match up well (at least MOST years) with #6 and #7 in the CCIW or #6-8 of the WIAC.

As Greg emphasized, conference strength means top-to bottom strength, with no extra weight given to tourney wins at the top. Virtually everyone gave the nod to WIAC over OAC as the top football conference because of depth (even before UWW became at least a near-peer of MUC).

Hugenerd · October 14, 2009, 08:24:39 PM

Here is some real data for you guys:

Massey 2006 (Overall Ranking)

4 - 4 tie

Spot 1 (CCIW): Ill Wes (319) - Carnegie (506)
Spot 2 (CCIW): Augustana (407) - WashU (568)
Spot 3 (CCIW): North Central (429) - Rochester (613)
Spot 4 (CCIW): Elmurst (525) - NYU (623)
Spot 5 (UAA): Chicago (633) - Carthage (695)
Spot 6 (UAA): Brandeis (742) - Wheaton (756)
Spot 7 (UAA): Emory (824) - Millikin (922)
Spot 8 (UAA): Case (849) - North Park (1136)

Massey 2007

4-4 Tie (WashU finishes 3rd in the country)

Spot 1 (UAA): WashU (392) - Augie (441)
Spot 2 (CCIW): Elmhurst (461) - Chicago (473)
Spot 3 (UAA): Brandeis (493) - Wheaton (502)
Spot 4 (UAA): NYU (506) - Carthage (533)
Spot 5 (UAA): Rochester (542) - North Central (587)
Spot 6 (CCIW): Ill Wesleyan (704) - CMU (792)
Spot 7 (CCIW): North Park (708) - Emory (897)
Spot 8 (CCIW): Millikin (883) - Case (1194)

Massey 2008:

UAA wins 7 - 1 (WashU wins national title)

Spot 1 (UAA): WashU (308) - Augustana (402)
Spot 2 (UAA): Brandeis (340) - Wheaton (473)
Spot 3 (UAA): Rochester (381) - Illinois Wes (577)
Spot 4 (UAA): Chicago (421) - Elmhurst (581)
Spot 5 (UAA): Carnegie (528) - Carthage (778)
Spot 6 (UAA): NYU (696) - North Park (792)
Spot 7 (UAA): Emory (782) - North Central (937)
Spot 8 (CCIW): Millikin (1095) - Case Western (1176)

Massey 2009

CCIW 7 - 1 (WashU wins title)

Spot 1 (UAA): WashU (228) - Wheaton (283)
Spot 2 (CCIW): Elmhurst (408) - Carnegie (620)
Spot 3 (CCIW): Augie (446) - Brandeis (731)
Spot 4 (CCIW): North Central (526) - Rochester (815)
Spot 5 (CCIW): Carthage (551) - NYU (940)
Spot 6 (CCIW): Millikin (569) - Chicago (1192)
Spot 7 (CCIW): Ill Wes (655) - Case (1229)
Spot 8 (CCIW): North Park (909) - Emory (1311)

As you can see from above, the conferences are very similar. They have essentially flip flopped each of the last 4 years. In 2006, CCIW won the top 4 matchups , then in 2007 the UAA won 4 of the 5 top matchups. Then in 2008 the UAA won the 7 top matchups, and then in 2009 the CCIW won 7 of the 8 matchups. If the trend follows, the UAA should have a stronger year this year than the CCIW.

According to the above results, I think the conferences are very close (maybe even too close to call), with a slight edge going to the UAA (obviously, in my opinion) because of recent performances in the NCAA tourney.

Hugenerd · October 14, 2009, 08:26:33 PM

In terms of the top-to-bottom argument. The UAA has had the stronger bottom of the conference (according to Massey) 2 of the last 4 years, and the CCIW has had the stronger bottom of the conference the other two years.

Mr. Ypsi · October 14, 2009, 08:47:23 PM

While I have not researched Massey's methodology, I strongly suspect that it highly weights tourney success, pulling up whole conferences (on the 'shirt-tails' of successful tourney teams). If I'm correct, the 2009 results are amazing for the CCIW, while the 2008 results would be expected.

How's that for rationalization?! But I suspect I'm right, however 'convenient' it may be!

Hugenerd · October 14, 2009, 08:57:44 PM

I am pretty sure Massey takes all the games into consideration. For example, even though 3 UAA teams made the tourney last year, it didnt really help two of them because they lost in the first couple of rounds. Obviously, wins over quality teams do help your ranking, but that should be expected with any system.

There is no "weighting" of the games near the end of the season. Here is the overview from the massey website:

"In essence, each game "connects" two teams via an equation. As more games are played, eventually each team is connected to every other team through some chain of games. When this happens, the system of equations is coupled and a computer is necessary to solve them simultaneously.

The ratings are totally interdependent, so that a team's rating is affected by games in which it didn't even play. The solution therefore effectively depends on an infinite chain of opponents, opponents' opponents, opponents' opponents' opponents, etc. The final ratings represent a state of equilibrium in which each team's rating is exactly balanced by its good and bad performances. "

Therefore, the only thing that the postseason helps you with is getting you more games (and potentially quality games). However, if you beat a bad AQ team in the first round and lose in the second round (or lose in the first round), the tourney can actually hurt you.

And I am a bit disappointed by your post Ypsi, without any research you immediately throw out excuses? I expected more. That is why I appreciate Sager's posts, at least they are well thought out, even if I do not always agree with them.

Mr. Ypsi · October 14, 2009, 09:24:05 PM

Hey, I already admitted it was 'rationalization'!

But your explanation actually confirmed my point - WashU's titles raised the entire UAA. Thus my (admittedly snarky) conclusion that the CCIW going 7-1 in 2009 was amazing, while the UAA's 7-1 in 2008 was (admittedly hyperbolically) only to 'be expected'!

Hugenerd · October 14, 2009, 09:43:24 PM

Quote from: Mr. Ypsi on October 14, 2009, 09:24:05 PM
Hey, I already admitted it was 'rationalization'!

But your explanation actually confirmed my point - WashU's titles raised the entire UAA. Thus my (admittedly snarky) conclusion that the CCIW going 7-1 in 2009 was amazing, while the UAA's 7-1 in 2008 was (admittedly hyperbolically) only to 'be expected'!

Well it technically also brings up the CCIW a little, because WashU plays games against the CCIW and plays in the same region (many common opponents-opponents, etc.). I dont really buy that excuse though, the differences in the top 7 are huge in 2008 in favor of the UAA (the UAA had 3 schools with a higher ranking than the first team in the CCIW). CWRU still has its expected rankings in the high 1100s - low 1200s, so it didnt help them very much. Also, the results of a given team are obviously more significant than opponents-opponents, etc.

This was my attempt to give you "real data" from an established algorithm, but you are still making excuses. It didnt even show a significant edge to either conference. To be honest, the post-season should have some weight, so the fact that you keep arguing that it shouldnt baffles me. If the top teams in your conference do well in the tourney, that makes your conference better, hence higher rankings. It doesnt make sense to ignore the post-season.

As I said before, the conferences have been very close over the past 3-4 years. You can make an argument either way. Nobody is going to gain any headway in this argument. There is no compelling evidence to show that either conference is supremely better than the other. I still contend that the UAA has the slight edge with the performance of their teams in the tourney in recent years