Herbie and the SEC love bug: wrecked by false logic about UM vs the SEC

Submitted by michelin on

I was disappointed yesterday and was tempted to believe what the UM-bashers were feeding us.  eg Herbstreit repeatedly  compared UM unfavorably to the SEC . Apparently, he thinks that this year's UM-Ala results can be compared to the "vastly superior" past performance of other SEC teams (eg in 2010 and earlier).  If you accept this arguable assumption, then other SEC teams should have done much better vs. Ala in 2010 than we did yesterday.  RIght?

Wrong.  UM this week did far better than Ala's SEC opponents did in 2010.,  Comaring the percentage of points scored by UM (vs Ala) with that of all Ala opponents in 2010, we had a record of  9-4.  More importantly, we went  7-2  vs the other SEC opponents.  In fact, our only SEC losses would have been a narrow 3 (percentage) pt loss to Ark and one loss to #1 ranked LSU, who we then would have beaten, in percent score, based on the national title game.

Granted, I am not sure whether we can compare this year's Ala team to last year's, but I am sure of one thing.  The logic of Herbie and the SEC love bug is a wreck.

Calculations

I computed the ratio of each team's score to that of Alabama's in each game (eg UM vs Ala 14-41 score means we scored  14/41=.34 or 34% as many points as Ala.  I computed this also for all Ala opponents in 2010 (based on the record given at the end), then I compared UM's % score with that of the other opponents.  Here were the results of these comparisons.

W 34-15 vs-KSt

L  34-41 at PSU

W 34-0  vs NoTx

L  34-37  vs Ark*

W 34-26 at Fla*

W 34=0) vs Vandy*

W 34-13 at Ole Miss*

W 34-16 vs  Tenn (.16)*

L 34-150  at LSU*

W 34-29 at Miss St*

L 34-47 Ga So

W 34-33  at Aub*

 

Nat title game

W 34-0 vs LSU*

 

*denotes SEC opponent

 

 

Here is the 2010 Ala record used in comparing UM vs. other Ala oppoents

 

ALA 2010 record

 

Opponent

   
   

Kent State Golden Flashes

Won 48-7

--

   

at Penn State Nittany Lions

Won 27-11

---

   

North Texas Mean Green

Won 41-0

---

   

Arkansas Razorbacks

Won 38-14

---

   

at Florida Gators

Won 38-10

---

   

Vanderbilt Commodores

Won 34-0

---

   

at Ole Miss Rebels

Won 52-7

---

   

Tennessee Volunteers

Won 37-6

---

     

---

---

   

LSU Tigers

Lost 9-6
(OT)

---

   

at Mississippi State Bulldogs

Won 24-7

---

   

Georgia Southern Eagles

Won 45-21

---

   

at Auburn Tigers

Won 42-14

---

 

2012 Allstate BCS National Championship Game

   

LSU Tigers

Won 21-0

---

           

 

 

 

 

 

 

 

http://www.fbschedules.com/ncaa-11/sec/2011-alabama-crimson-tide-footba…

PurpleStuff

September 2nd, 2012 at 6:20 PM ^

The fall of FSU and Miami has pushed more talent to the SEC, but the best teams from that region and from the west coast have more often than not been better than the best teams from the Big Ten.  Our '91 team steamrolled the conference schedule, yet gave up 51 to FSU in the Big House and lost 34-14 to Washington in Pasadena.  SC and Oregon are 4-0 against us in the last decade.  OSU's SEC struggles are well documented.  We all know Bo's bowl record.

It isn't a universal fact (PSU beat Miami in the mid-80's, OSU did it in 2002), but I'd usually bet on the top dog from that region over the best in the midwest.

 

befuggled

September 2nd, 2012 at 8:25 PM ^

Grbac unfortunately had two interceptions returned for touchdowns. The second one was on a desperation fourth-down play when we still had an outside chance of winning the game. Given that we had the ball in the red zone four times that day without scoring (thanks to Grbac's other two interceptions and turning the ball over on downs twice), I feel comfortable in saying that the game was closer than the final score indicated.

The Rose Bowl game against Washington, though, was probably worse than the final score. I thought we were lucky to be down by only 6 at the half, and our first touchdown was something of a gift.

TheLastHarbaugh

September 2nd, 2012 at 6:20 PM ^

I thought after the game, despite how awful things were, that we could make an argument as being the 2nd best team in the SEC.

I think Alabama is that much better than everyone else

turtleboy

September 2nd, 2012 at 9:00 PM ^

I thought so, too. When the crowd started cheering SEC-SEC-SEC I was confused because right now the SEC (west)  has 1 great team, and another good team with a near-great defense. The rest are pretty average and have no answer for Bama/LSU. Tennessee, Kentucky, Vanderbilt, and Florida are a mess, South Carolina has 3 great players, but aren't a great team,  who knows whats happening with Auburn but they've stepped way back since winning the Cam Newton lottery, and Georgia was great with Stafford, but they've lost almost every major in-state recruiting battle for years, now. Michigan would win the East.

Aspyr

September 2nd, 2012 at 6:22 PM ^

They did mention that we had at least crossed the 50 (and scored) before half time something that LSU hadn't done in the national championship game. FWTW

But the fact of the matter is that even though we have historically done well against the SEC we have been curb-stomped in the last two meetings. As soon as we start beating top ranked teams again all the love will come. The press and ESPN for that matter love winners. 

turd ferguson

September 2nd, 2012 at 6:24 PM ^

I respect the effort, but that's a very questionable metric. For example, I'd have felt much better with a 10-3 loss last night, but I think you're saying that would have been a bigger blowout than 41-14.

The reality is that the elite Big Ten teams don't stack up to the elite SEC teams right now. As you go down the standings and compare (mid- and lower-level teams), I think the Big Ten is more competitive, but there's no real disputing the disparity up top.

michelin

September 2nd, 2012 at 7:46 PM ^

.your example of a 10-3 score points out that my metric using ratios does not make as much sense if the numbers are very small.  However, that only applied to one Ala game in 2010--the 9-6 LSU win.  I already gave LSU credit for that win, so any distortion due to the metric I used is irrelevant.  It would not change the results.

It is true that, if we used only point spread, the results would change---we still would have a respectable record, just not as good.  However, for reasons stated above, I feel that the point spread is not a good measure--particulary when the scores are larger---as was the case for all but one 2010 Ala game.

turd ferguson

September 2nd, 2012 at 8:50 PM ^

For what it's worth, my view on this is that if a model doesn't make sense in some regions of data, it probably isn't great in other regions of data. 

Here's another example of where this produces strange results.  The key result for you is that our point total was 34% of Alabama's point total.  Let's say that the Gardner touchdown never happened, so we only scored 17% of Alabama's total.  Would that have felt like a completely different outcome to you, indicative of a much bigger blowout?  It wouldn't have to me. 

(Side note: the issue here is another "small numbers" thing, but this time it's 1 TD vs. 2 TDs.  I think those problems are inescapable in analysis like this.)

michelin

September 2nd, 2012 at 9:38 PM ^

Take your example a step further.  Would it not also be a much worse outcome if we didn't score the other TD either and got shut out?  It's only seven more points, if we use point spreads, but a shutout should count for something more.. 

I am not saying that my measure is the only valid perspective---far from it.  IMO the only way to have a truly normative measure would be to calculate the probability (Pr) of winnng the game at different score combinations eg eg Pr=.01 at 21-0 vs Pr=.015 at 41-14 vs Pr=.005 at 41-7.  Of course, these probabilities would depend on the stage of the game (eg halftime vs end of 3rd quarter) and they would converge to zero as the game wound down till the end.  Yet, it might be shown that one series has a consistently higher probability over a reasonable range of different times left on the clock.

GIven the difficulty of doing such an analysis, one could use a variety of different measures.  Using either mine or the one you suggested, the results disconfirm the pundits' idea:  that UM isn't as good as SEC teams (other than Ala).

wolverine1987

September 3rd, 2012 at 10:49 AM ^

That we aren't as bad as the rest of the SEC (outside Bama and LSU) as people thing we are? That is really not much of an argument, and looks to me like a severe over-reaching for the slightest glimmer of positivity. That's ok I guess, but it doesn't really prove anything.

I think it's now very clear, to both objective and non-objective viewers alike, that the SEC praise from the media the last few years is, as tempting as it was to believe otherwise--very true.

michelin

September 3rd, 2012 at 12:52 PM ^

the intensity of the Michigan-bashing by news commentators, like Herbie, was a distortion of the truth.  Alabama has been beating everybody that badly--and typically even worse.

You can make an argument from evidence entirely independent of this game, that the B10 is not as good as the SEC.  But you cannot generalize from this one game and say the UM is not as good as SEC teams (as a group).   A day after I posted this, Brian also noted the poor record of other SEC teams vs Ala (as well as MSU's 2011 bowl game vs Ala, which they lost by about two touchdowns more than we did).

wolverine1987

September 3rd, 2012 at 5:41 PM ^

that "Michigan is not yet where they want to be in terms of returning to the elite (paraphrasing), but after a couple more of Hoke's excellent recruiting clases they will be. This would be a much more competiitve game in two years." Which in my opinion, is an undeniably true statement, both for now and for later.

I agree that we are better than many SEC teams, but this much is true: as one of the (let's say) 3-4 elite teams in the B1G, we do not stand up to the 3-4 best teams in the SEC, as I think the result against LSU would be similar. And both of the top two SEC teams are WAY better than the top couple B!G teams.

michelin

September 2nd, 2012 at 7:05 PM ^

You can point to some examples like that.  But Is a 21-0 deficit (needing 3tds and a fg to win) really better than a 41-14 score (needing four tds to win)?  In the former case, a team that has not scored a single point needs four scores to win  In the latter case, a team that has already scored 2tds needs four more scores.  Depending on the time left, I would usually bet on the latter.

SysMark

September 2nd, 2012 at 10:52 PM ^

You're right and I generally dislike this type of analysis for the reasons you mentioned.  However in this case I do think his overall point is valid - compared to what the other SEC teams did vs. Alabama last year we look at least as good as most of them.  And that includes LSU to a degree - they were shutout 21-0 and didn't cross midfield in one game.

Sten Carlson

September 2nd, 2012 at 6:40 PM ^

I am all for giving Alabama their due -- they're an amazingly well coached and talented football team.  But, I cannot shake the feeling that they SHOULD be that good given the built in advantages that they have over the rest the other non-SEC schools.  In fairness, I won't even bring up the illicit advantages because I have no real proof that they occur -- although I believe that they do.

This is not "sour grapes" Bama is great, and Michigan is not even close to their caliber, and maybe never was nor will be.  But, just like I cannot take MLB pundits serious when the fawn over the NY Yankees who neglect to mention that TOTALLY unfair playing field the Yankees play on due to their wealth, I cannot take a CFB pundit serious when he fawns over Bama without mentioning oversigning.

On www.oversigning.com Alabama was +12, while Michigan was only +1 in their "Oversigning Cup Points."  I think this practice has been curbed significantly, which might help Michigan get back onto par with programs like Alabama, but the effects are quite obvious, though not talked about by the fawning commentators.  It's a lot easier to build that ever so valuable depth when your program effectively gets an extra full recruiting class every four years.  Compare that the horrible attrition that Michigan went through from 2005 until Hoke stopped the bleeding, and you see the disparity that was on display last night.  Again, cudos to Bama, you're bad ass.  But, why not discuss this as a factor?

somewittyname

September 2nd, 2012 at 6:42 PM ^

If you just ignore Alabama and LSU then the conferences are equal? That's like saying consider the B12 but ignore OU and Texas. Or the Pac12 and ignore USC and Oregon. There is absolutely no doubt that the SEC is considerably stronger than the B10.

michelin

September 2nd, 2012 at 7:02 PM ^

We went 1-1 vs LSU in my analysis but LSU essentially had home field advantage for both games, so you could argue that we came out ahead of them.

I agree that we are not in any way as good as Ala.  Yet, the pundits do not stop there: they conclude we are not in the same class as the SEC (made up of many other teams).  I do not think that last night's game proved that.  If it proved anything, it is the opposite.

club2230

September 2nd, 2012 at 7:26 PM ^

The dominance of the SEC is really distorted by the fact that the best of the SEC is clearly better than the best of the rest.  Sure, Auburn won a title recently, but they would not be in the top 2 of the Big Ten this year.  Florida is the same too.  Arkansas has never won anything that I can remember as being significant, nor has South Carolina.  Georgia hasn't won anything either although their history makes me think they have been better than they have been recently.

If Alabama and LSU left and formed their own conference then they would be the ones chanting S-E-C and the former members of the conference would be relatively quiet.  

We'll get there with Hoke focusing on the Big Ten first and formost.   We need to become a consistent midwest power before we can expect to be title contenders year in and year out.  We even have an advantage in the midwest as there are now only two teams that can legitimately vie for the top recruits enmasse.  That is not the case in the south where there is FU, FSU, Miami, Ala, LSU, and Auburn to contend for the southeast recruits.

NateVolk

September 2nd, 2012 at 8:16 PM ^

I do remember that one of the hang ups to Texas joining the SEC was their demand that the academic standards for football players be equalized among member schools. The request was specifically aimed at Alabama.  

Alabama was having none of it.

LSAClassOf2000

September 2nd, 2012 at 8:50 PM ^

Sort of taking this in a different direction and trying to build on it a little, using the 2011 schedules for Michigan and Alabama, I figured out the same ratio for each game in the schedule (using "0" for shutouts and losses, for purposes of this analysis), and got the average for each team. It comes to 0.46 for Michigan and 0.3 for Alabama. 

If you adjust the points against us downward to "simulate" the sort of luck teams had against Alabama offensively last year (34% less scoring on average), we actually still lose to MSU and the Iowa game becomes a statistical tossup. Conversely, if you "simulate" the sort of offense that Alabama's opponents put up against up (i.e., 0.46/0.3 this time), then Alabama's record is unaffected. That being said, we averaged 34 points per game last year, and Alabama average 36, so so it may stand to reason that records would remain largely untouched, but it's interesting to think about the sorts of margins both teams were winning by - we averaged 24 points, whereas Alabama averaged 30 points in their wins. 

JoJoBlue

September 2nd, 2012 at 10:55 PM ^

Lets be honest:

We aren't in the the same field with the likes of Alabama.  I somewhat expected the score last night, given our performance last year and given Nick Saban's propensity for having his teams coached extremely well...and to (cough) oversign with the best of them.

 

But I digress.  Making broad generalizations about the B1G and SEC from this 1 game is a little silly.  The SEC is a solid conference, but essentially boils down to two elite teams, a couple decent teams, and some perrenial bottom feeders.  I don't see how this is much different from the PAC-12.  Arguably, last year, the PAC-12 was a better conference (all around) with the likes of USC, Stanford, and Oregon, while the SEC had LSU and Alabama.

 

As for the B1G, I see us as having more solid teams, but no real elite teams, right now.

Tater

September 3rd, 2012 at 7:45 AM ^

The Big Ten is almost as deep as the SEC now, but Bama and LSU are still enjoying the advantages of five years of over-recruiting; that skews the relative strengths of the conferences.  Bama is going to do the same thing they did to Michigan to a lot of top-level SEC teams.  

As much as I hate to say it, we almost have to cheer for Sparty to remain a good team, becuase the NCAA gutted Penn State for the next ten years, and that negates the acquisition of Nebraska.  

Maybe the next time Alabama gets caught cheating, the NCAA will gut them for ten years instead of giving them "punishment" that amounts to fifteen minutes of "time out."