Wednesday, January 03, 2007

Bowl Challenge Update

After last night's Orange Bowl -

Big East 1st tier (Louisville) over ACC 1st tier (Wake Forest) - +3 Big East

Current Standings -

ACC = 11

Big 10 = 5

Big 12 = 6

Big East = 9

Pac 10 = 5

SEC = 13

In our last update, the comment section of our Bowl Challenge had some questions about the methodology being used. What I was trying to achieve in the contest was a way of comparing conferences for overall conference strength. Virtually every BCS conference has one or two top tier teams that can compete with any team in the nation. In the Pac Ten, for instance, USC and Cal have shown they can be competitive with anyone. However, it remains my belief that a conference is strong only if its middle tiers are competitive. In the case of the Pac 10, after USC and Cal, I do not believe many of the remaining teams would be anything but bottom tier, non bowl teams in most other conferences. A USC gets a great advantage from having to play in a conference of weaker teams all year long – an advantage in terms of injuries, win-loss records and general wear and tear.

What I was trying to devise was a way to compare the different levels of teams between conferences. For instance, I do not think if Florida is fortunate enough to beat Ohio State that it proves the SEC is the best conference. Rather, it proves 1 SEC team is better than 1 Big Ten team. So, in devising my point system, I tried to make it “neutral” for games between teams of the same standing in different conferences.

If anything, this methodology worked to the advantage of the ACC in the Florida State – UCLA match up. If you look at the Pac Ten and the ACC, I think you could fairly rank these teams in their conferences as follows –

UCLA 3rd best Pac 10 (After USC and Cal).

FSU 6th best ACC team (After Wake Forest, Virginia Tech, Boston College, Georgia Tech and Clemson).

So, UCLA went into the 2nd tier Pac 10, and FSU the 3rd tier ACC. By beating UCLA, FSU got 4 points, 1 point more than they would have gotten had they beaten a 3rd tier team.

There was some criticism of my ranking Tennessee a 3rd tier SEC team, and Penn State a 2nd tier Big 10 team, but I think those rankings are right on.

First the SEC (prior to bowls)

1. Florida 12-1
2. LSU 10-2
3. Auburn 10-2
4. Arkansas 10-3
5. Tennessee 9-3

It is hard to argue that UT is not the 5th best team in the SEC. They are ranked 5th best in the SEC by the BCS, which I used as a default to rank teams where there was a question.

Now the Big 10 –

1. Ohio State 12-0
2. Michigan 11-1
3. Wisconsin 11-1
4. Penn State 8-4
5. Purdue 8-5

I also think, based on the BCS rankings, and simple logic, that Penn State is the 4th best team in the Big 10.

So, we had the 2nd tier Big 10 team against the 3rd tier SEC team.

And that is exactly the comparison I am trying to make – a comparison between conferences of teams similarly situated (or not) in their conference. Who has the deepest conference strength?

In regard to non-BCS conferences, I think there is a misunderstanding. Non BCS teams only have two ways of scoring - +2 if you win, -1 if you lose. I chose this because I did not think it appropriate that a team get even as many as the “neutral” 3 points for beating a non-BCS team, and that there should be a penalty for losing to one. While Houston was called a 1st tier team Conference USA team, South Carolina still got only 2 points for beating them (as opposed to the 5 they would have gotten for a 3rd tier team beating a BCS 1st tier team).

It was difficult to know how to score the non-BCS conference teams, but with so many playing, you couldn’t ignore them. The Pac 10 played 2 of their 6 bowls against non-BCS, and lost to both. If they didn’t count at all, the Pac 10 would be getting credit they did not deserve.

I also am strongly against the systems used by networks like ESPN that merely rank wins and losses. Take the Cal – Texas A&M game. Cal should – and did – beat Texas A&M. Why should Cal get equal credit (1 win) by beating an obviously inferior Big 12 team? Cal is clearly the 2nd best Pac 10 team. If they played the 2nd best Big 12 team, it would have been Texas. Merely assigning wins-losses doesn’t tell the whole picture.

Plus, using a percentage system is also flawed, and cuts against conferences with many bowl teams. A conference with 3 teams that goes 2-1 would be considered the same as a conference with 9 teams that goes 6-3. Or, a conference with 8 teams that goes 4-4 considered inferior to a conference with 3 teams that goes 2-1. Percentages don’t make sense because they cannot be evenly applied – some conferences have even numbers of teams, some odd. You can’t get .500 with 3 teams, for instance. I think conferences like the ACC or SEC with 8 and 9 teams deserve some credit for getting that many to bowl games.

It looks now as if the SEC will win, but that was not a foregone conclusion at the outset. The SEC still had to win games to score points, and they have.

Anyway, remember, this is an inaugural contest, and at the onset, I solicited comments for ways to improve the format. I continue to solicit you input, so that we might perfect the methodology next year.

6 comments:

D. Tensor said...

Why use BCS rankings to determine order within a conference and not the conference standings? (Does this mean that the SEC championship game is a sham as it actually matched the #1 and #4 teams in the conference rather than #1 and #2 as I would have expected?) - or - alternatively, calculate points using both BCS and conference standings and use the difference as an estimate of potential error.

I think some factor ought to be given to size of the conference - perhaps divide the final value by number of teams in the conference - otherwise it seems grossly unfair to the Big East and slightly less so to the Pac Ten (very slightly so for Big Ten). Try reversing the system and count losses instead of wins according to the same system (reversed) and I suspect that larger conferences will also score more points in the loss department. It depends somewhat on definition of conference strength. My own feeling is that a small conference can be as "strong" as a big conference. If, for example, the Pac-ten/Big-ten/Big East/ACC magically merged next year the resulting super-conference would score a lot of points in this type of system which only counts wins. It would not mean that the super-conference was "stronger" - just that it would have lots of good teams (but also lots of bad teams).

Might be useful to treat independents and non-BCS in the same way - not clear to me why Navy gets their ranking. Actually, making ND tier 1 is debatable - they were spanked by tier 1 teams even prior to their bowl and had a BCS rank below some SEC tier 2 schools. Penalties seem odd - losing to a Boise State does not seem that bad to me to deserve a penalty as that team would likely play at least near the .500 level in a BCS conference (and is certainly better than Navy) Perhaps call all non-BCS/independent schools with BCS rank 1-10 tier one [or 1-11 if you really think ND is top shelf], rank 11-20 tier two, and the rest tier three.
Anything below 50 gets penalty points.

I don't think these changes will make much difference in the final points but will make it appear less arbitrary (and consequently more believable, at least to non-SECer's)

Henry Gomez said...

Tensor,

You should read the comments on the previous post about this challenge. They should give you some insight as to why the scoring is the way it is. But let me address some of your points.

Why is LSU in the first tier and not Arkansas? Well LSU beat Arkansas for one. LSU was the at large BCS team from the SEC not Arkansas. Yes Arkansas won the division but Arkansas played a different schedule than LSU. They didn't play Florida, for example, until after they had won their division. I think it's pretty clear cut that LSU is better than Arkansas.

In Confs with 2 divisions, the standings aren't instructive because one division may have the top 3 teams.

ND gets treated as a tier 1, because their ranking and bowl standing indicated that. The purpose of the exercise is to see if in fact these teams are as good as advertised.

The tiers for each conference were not relative to teams outside of their conference. In other words we go into it assuming the 3rd best team in the Pac 10 is the as good as the 3rd best team in the Big 10. It doesn't matter that we think the 3rd best team in the Big 10 is much better than the 3rd Best team in the Pac 10 ten. That's what we're trying to find out.

Each conference got 2 tier one teams and 2 tier 2 teams. These are the top 4 teams in their conference. Since the tiers are relative only to the other teams in the conference you can't say "there are no tier 1 teams in the X conference." There's always a top, middle and bottom.

You get penalized for losing to a Non-BCS conference team because BCS teams are supposed to be superior. We can look back and say that losing to Boise is no sin but that's using hindsight. An Oklahoma beating of BSU would not have surprised anybody.

Anyway it's not a perfect system but neither is the one they foist upon us in that Bowl Challenge cup.

Henry Gomez said...

To clarify one point...

Mergz challenge format makes a couple of assumptions going into the challenge:

1. All BCS conferences are equal. This assumption is what gets tested with the challenge.

2. All BCS conferences are superior to non-BCS conferences. The reason is because we are trying to measure the strength of the BCS conferences and not the non-BCS conferences. The assumption is that the Big XII champ (or the SEC or the Big 10 champ) should beat the Mt. West Champ otherwise the Mt. West should be a BCS conference too. If the BCS team that plays the Mt. West champ loses then we can say that the conference is not as strong as it is billed to be because it couldn't beat the champ of a supposedly lesser conference. Again we don't create the assumption but it's out there and we accept it to test it.

d.tensor said...

Thanks for the time and response. I think it is an interesting test and in the right direction. My comments were intended primarily to make it more robust (at least in my Big Ten/Mt. West biased opinion).

Actually, I had read the other posts but the rational for the precise independent's ranking was unclear.

I would argue that while non-BCS conferences are weaker, the champion is not necessarily so.
Conference strength is not related to champion strength [see the 1984 BYU team]. (just as Florida defeating OSU does not automatically prove the SEC, overall, is better than the Big Ten, as is stated somewhere).
---------------
[actually, now it bothers me that Arkansas was the West champion but is the third-best team in the West - does this mean the BCS system is more accurate than a playoff system?]
[I agree other conferences are also flawed (e.g. Big Ten) but I honestly thought conference championships were a good idea -
suppose Arkansas had squeaked by Florida in the championship - who would be the best team in the SEC?]

Henry Gomez said...

Tensor,

I agree that the Mt. West champ could in fact be better than the Big XII champ. In fact I would argue that that's exactly what happened. BSU proved it was the better team. They allowed OK to get back into a game they pretty much had sealed up. But the theory (not one I subscribe to) is that the Mt. West Champion should never be superior to the Champion of a superior conference. It's a theory that on its face is ridiculous. The best team in the country could in fact play in the weakest conference. That's unlikely however because they will not have been tested in a close games. But that's a exactly why a playoff is required. The Gators played 13 games and 9 of them were against conference foes. 1 game against an in-state non-conference rival that was down this year (FSU) and then 3 patsies (1 is actually a Bowl team, Southern Miss and one of the others was a Bowl team last year, UCF). But with so much intraconference play and so little interconference play it's hard to judge which conference is good from year to year.

Any time you have divisional play (like you do in the SEC) there's a chance that one of the division runner ups will be better than the winner of the other division. Take baseball for example. Last year St. Louis had the 13th best record in baseball and won the world series. That's because they played in the weakest division in baseball. They got hot at the right time. They probably would not have won the World series if the playoffs were more like the regular season (long). But in a short series anything can happen. There is NO way that St. Louis was the best team in baseball but there's no arguing that they are the champs.

Your question of who would be the best team in the SEC if Arkansas snuck past Florida is the root of all this debate about playoffs.

My answer is that is that Arkansas would have earned the title of best team (just like the cardinals) in the SEC regardless of what our personal feelings were. They would have been the best team in their division and would have beaten the champion of the other division.

But right now CFB doesn't work that way on the national level. Example: LSU killed Notre Dame. They were the 4th ranked team in the country going into their bowl. Michigan lost, they were ranked 3rd. If Florida were to lose to OSU I can almost guarantee you that LSU would jump Florida in the final AP poll. This despite the fact that Florida beat LSU 23 10.

What sense does that make? So I'd rather take my chances on a cinderella winning an 8 or 16-team tournament for a championship which is more legitimate than the current system which is merely a two-team playoff.

Anonymous said...

Thanks again. Good arguement. I think the Wall Street Journal had an article ranking teams/conferences using some scheme based on draft picks and number of pro players (and sucess as pros) over the last few years. The number one team (by quite a bit) was Florida State, but the SEC had the most number of teams in the top ten as I recall. Talent alone doesn't always win but it helps. It's about the only other way to judge teams.

To be honest and since no system is perfect I rather like the multiple championship bit (e.g. LSU is happy, USC is happy) but whatever. I'd be most worried about cirrhosis of the liver if I had to go through a playoff series every year.

ps. Boise is in the WAC, not Mt. West. My fault, I think.