Top Conferences Top-to-Bottom Using Sagarin Predictor

[banditbdl]

The problem with the using the Sagarin for this is that such a large part of it is based on Strength of Schedule which is a pretty meaningless measure in college football since there are so few non-conference games and the vast majority of them are bought and paid for by schools from the big conferences.

9 of the top 10 teams for Strength of Schedule according to Sagarin are from the Big 10. The other two Big Ten teams, Iowa and Wisconsin are ranked 13 and 18 respectively. So you figure maybe the Big 10 dominated its non-conference games huh? Well, sort of.

Here was the Big 10s record against some of the other conferences:

W L
Big 12 0 2

ACC 1 0

Pac-10 1 2

SEC 1 0

NotreDame 1 2

Big 10 was an underwhelming 4-6 against the more respected opposition. They fared a lot better against the rest though going a combined 16-0 against the M/WAC and 3-0 against the dying Big East with all but 2 or 3 of those 19 games being played at Big 10 stadiums.

All the Sagarin and your rankings have shown is that the Big 10 is indeed awesome at scheduling patsies at home and not screwing them up. But we all knew that already.

[lastchance]

Notre dame 1-2 is completely fine, considering Purdue, MSU, and Michigan aren't the class of the Big Ten, and Notre Dame is a top 10 team.

OSU loses to Texas, and that's fine, because Texas is the second best team in the country.

ISU (4th best in Big 12) (a decent team) beat Iowa (6th best in Big Ten).

Illinois (worst team in Big Ten) loses to California (4th best in Pac-10).

ASU (5th best in Pac-10) beats Northwestern (7th best in Big Ten).

All of these losses are a superior team beating a worse one. The Big Ten's cream of the crop hasn't lost to any bad teams, and they've beaten some decent ones. Of those 6 losses to teams in major conferences, 1's against Texas, 2 are against Notre Dame, another is because of Illinois, and the others are lesser teams from the Big Ten against good teams out of conference.

The evidence is not damning.

[TomCollins]

I'm going to add the Sagarin rankings for each of these matchups.

[ QUOTE ]

Here was the Big 10s record against some of the other conferences:

W L
Big 12 0 2

Big 10 #1 vs Big 12 #1 (Texas vs. OSU) L
Big 10 #6 vs Big 12 #4 (Iowa vs Iowa St.) L

ACC 1 0
Big 10 #7 vs. ACC #9 (Wisconsin vs. NC) W

Pac-10 1 2
Big 10 #11 vs. Pac-10 #5 (Illinois vs. Cal) L
Big 10 #9 vs. Pac-10 #4 (NW vs. ASU) L
Big 10 #8 vs. Pac-10 #8 (Purdue vs Arizona) W

SEC 1 0
Big 10 #10 vs. SEC #11 (Indiana vs. Kentucky) W

NotreDame 1 2
Big 10 #3 (Michigan) L
Big 10 #5 (Michigan St) W
Big 10 #8 (Purude) L


[/ QUOTE ]

So the better teams of some of the top conferences are better than some of the weaker teams of the Big 10. And Texas beat OSU in a close game.

Doesn't every major conference schedule patsy non-conference games? But I guess the MAC Champion is just as much a patsy as Rice or Louisiana Lafayette.

[banditbdl]

The fact that your Predictor has a 2-6 Mich. St. ahead of Wisconsin, NW, and Iowa is evidence enough that it is deeply flawed. Not to mention that it rates Minnesota ahead of Iowa after getting completely blown of the field by the Hawkeyes on Saturday. Just because the Sagarin might be the best ranking out there doesn't mean it still doesn't suck.

And yes, all the conferences schedule the patsy games which as I said makes the sample size to judge these conferences against each other way, way too small.

The Big 10 may just be the best conference this year, but you can't prove it using a computer ranking with such an inadequate sample size.

[MCS]

[ QUOTE ]
The fact that your Predictor has a 2-6 Mich. St. ahead of Wisconsin, NW, and Iowa is evidence enough that it is deeply flawed.

[/ QUOTE ]

No it's not. What it means is that you have to look beyond just win-loss record. You have to look at who a team played and how close their games were.

For example, one of MSU's out-of-conference games was a win over Notre Dame. They went to OT with Michigan and played at Ohio State and against Penn State to approximately 10-point losses. They destroyed the bad teams they played. That sounds like the profile of a decent team.

[ QUOTE ]
Not to mention that it rates Minnesota ahead of Iowa after getting completely blown of the field by the Hawkeyes on Saturday.

[/ QUOTE ]

So what? This type of "conflict" is unavoidable when you rank teams. Georgia Tech beat Miami, who beat Virginia Tech, who beat Georgia Tech, who beat Miami...

If the predictor is so bad, why do experts use it to predict games?

[cokehead]

[ QUOTE ]
So you figure maybe the Big 10 dominated its non-conference games huh?

[/ QUOTE ]

But the Big Ten atleast played some games against more respected opponents. look at the SEC, who have any of them played? LSU barely beat ASU, and Tenn got blown out by ND. What other games have there been?

[Jack of Arcades]

[ QUOTE ]
LSU barely beat ASU

[/ QUOTE ]

On the road and not long after Hurricane Katrina. Let's not forget that ASU kept USC to 38 as well.

[banditbdl]

[ QUOTE ]
If the predictor is so bad, why do experts use it to predict games?

[/ QUOTE ]

I'd like to see an expert who would set Mich. St. as a point spread favorite against any of the three teams I mentioned.

I find it funny that on a poker board where people talk about needing tens of thousands of hands to have a relevant sample size people are willing to accept a computer ranking to compare these conferences based on a handful or two of interconference games.

The simple fact is we don't have even close to enough quantitative information to say one football conference is better than the others.

[TheNoodleMan]

I'd ask Jeff Sagarin about it myself, but he seesm obsessed with time zones lately. I have seen him at the courthouse a few times in the last month arguing that Indiana should be on central time. It is pretty sad to watch a bunch of politicians ignore the smartest person in the room.
He has a strange web page befitting a statistical guru http://www.sagarin.com/
His sports stuff can be found here. <font color="yellow"> </font>

[MCS]

[ QUOTE ]
[ QUOTE ]
If the predictor is so bad, why do experts use it to predict games?

[/ QUOTE ]

I'd like to see an expert who would set Mich. St. as a point spread favorite against any of the three teams I mentioned.

[/ QUOTE ]

MSU is 0.05 ahead of Iowa.
MSU is 1.31 ahead of Wisconsin.
MSU is 4.94 ahead of Northwestern.

So what this says is that MSU, Iowa, and Wisconsin are all pretty close, and that all three are clearly better than Northwestern, but not by a ton. None of this strikes me as incorrect.

EDIT: I do compare predictor numbers to the actual lines, and they're almost always very close. So I think the line would be about 0 against the first two teams, and about 4 against Northwestern. Maybe Iowa or Wisconsin would be favored by a point or two because of public perception. Any more than that, and I'm on MSU.

When Alabama was playing LSU, everyone thought Bama was such a steal as a home dog. Predictor had LSU favored by about 2.