Michigan State is Ours . . .

[MichiganDad3]

Not to be debbie downer or anything (I know, shocking that tschu would be Buzz Killington), but if Nebraska truly is a top-10 team or a team capable of running the table, we just don't have any data that backs that up right now. Maybe we'll generate some of that awesome data Saturday.

I agree that there is not sufficient data to prove NU is top 10, but there is also not sufficient data to prove we are not top 10. And where is the data supporting MSU is top 10? A beatdown from Oregon, and three wins over Sam Sargent Head State University for athletically challenged.

 

[MichiganDad3]

Who has nebraska beaten? Nobody. Michigan state was in control most of the Oregon game. It will be a close one and can go either way

Miami and Illinois are better than any team MSU has beaten so far. I think NU wins easily if the turnovers are even or in our favor. I would also like the NU OL to cut the penalties in half.

They had Oregon on the ropes, until the 3rd quarter when they blew the lead.

Thank goodness the MSU NU game is 4 full quarters. It looks like MSU is world beaters for one half of a game

 

[tschu]

Not to be debbie downer or anything (I know, shocking that tschu would be Buzz Killington), but if Nebraska truly is a top-10 team or a team capable of running the table, we just don't have any data that backs that up right now. Maybe we'll generate some of that awesome data Saturday.

I agree that there is not sufficient data to prove NU is top 10, but there is also not sufficient data to prove we are not top 10. And where is the data supporting MSU is top 10? A beatdown from Oregon, and three wins over Sam Sargent Head State University for athletically challenged.

They did not get beatdown by Oregon; in fact they dominated half of the game. And in fact that result was certainly within the range that one would expect from that matchup on the road given that Oregon is very highly rated. It was even closer if you discount that final garbage TD scored as Oregon was running the clock out, but the computers don't do that.

Like I said, all that the models do is take all available information into account and fit the best possible ratings to that data - the ones that show the least overall error. Since there's only 5 games worth of data, of course the margin for error is going to be significant. But all I can do is point out the best information that's available at this point.

 

[HUSKER FREAK]

I would take a Tommy Armstrong ran team every time over a T-Mart ran team. T-Mart was one hell of an athlete could toss the ball decent at times. What I hated about the T-Mart ran team is that when he got hurt we were done. We put all of our eggs in his basket and it back fired. This would have been the same when Crouch was running the offense he was an outstanding Athlete and stayed healthy so he turned out how T-Mart could have, but that is a big risk to take. Tommy is tough a decent runner he can toss the ball pretty well and will only get better but he has the respect of his team mates and is confident on and off the field.

 

[Saunders]

Here's some quick and dirty numbers.

Big 10 Defense - Overall

Big 10 Defense - FBS Only

Big Ten Offense - Overall

Big Ten Offense - FBS only

 

[QMany]

I love that our rankings improve with FBS only.

 

[MichiganDad3]

Not to be debbie downer or anything (I know, shocking that tschu would be Buzz Killington), but if Nebraska truly is a top-10 team or a team capable of running the table, we just don't have any data that backs that up right now. Maybe we'll generate some of that awesome data Saturday.

I agree that there is not sufficient data to prove NU is top 10, but there is also not sufficient data to prove we are not top 10. And where is the data supporting MSU is top 10? A beatdown from Oregon, and three wins over Sam Sargent Head State University for athletically challenged.

They did not get beatdown by Oregon; in fact they dominated half of the game. And in fact that result was certainly within the range that one would expect from that matchup on the road given that Oregon is very highly rated. It was even closer if you discount that final garbage TD scored as Oregon was running the clock out, but the computers don't do that.

Like I said, all that the models do is take all available information into account and fit the best possible ratings to that data - the ones that show the least overall error. Since there's only 5 games worth of data, of course the margin for error is going to be significant. But all I can do is point out the best information that's available at this point.

A top 10 team should not get beaten by 19 points. I am not impressed with MSU. I am still on the fence about NU. We have been impressive the last two games, but we have a lot of injuries, and we have not played any top tier competition.

 

[Landlord]

A top 10 team should not get beaten by 19 points.

According to...what?

 

[gossamorharpy]

A top 10 team should not get beaten by 19 points.

According to...what?

Have a feeling that MichiganDad3 purely saw the scoreline or tuned in for the last few minutes. That was a back and forth, one possession game until midway through the 4th, on the road; in one of the toughest game day environments to play in, against a higher ranked team who has the most impressive win to date of any top ranked title contender... MSU may have lost by 19 points but that game was much closer than the final score indicates, a few of my oregon buddies weren't feeling too confident there until the end.

 

[WhatDoIKnow]

The MSU D got worn out in the second half by Oregon. Which is understanable having to chase those guys all over. It would be nice if we could wear them out with power football.

 

[Guy Chamberlin]

I do hope some folks here will stop confusing optimism with naivete.

 

[Danny Bateman]

Here's some quick and dirty numbers.

Big 10 Defense - Overall

Big 10 Defense - FBS Only

Big Ten Offense - Overall

Big Ten Offense - FBS only

You take that FCS game away, and there's a 26 yard gap between defensive yards allowed MSU vs. NU.

At the same time, they average 20 less YPG offensively outside of that FCS beatdown. They laid about 600 of their total yards (or roughly 30% of their season output) on that perennial powerhouse, Jacksonville State.

Awful lot of excuses being made for Sparty as to why they lost to Oregon around here...

 

[irafreak]

I think you could argue that computer rankings may be biased since they are designed by people. I mean why should strength of schedule be factored in but they pull out the margin of victory? Is it absolute fact that margin of victory means nothing? To the guy/group that wrote the program it is...

What? Are you referring to the BCS computers? Because those are obviously horrible. Any decent model whose goal is to predict results should include margin of victory. The fact that the BCS took margin of victory out of the equation was embarrassing.

I don't look at sh**ty models like that. Sagarin was a BCS computer, but he had to modify his formula and make an entirely new ranking, separate from the ones he posts each week on USA today, to give to the BCS. His normal rankings include margin of victory, and actually the majority of it is determined by margin of victory. But he had to take it out of the equation to appease the BCS. And it was hilarious how bad and how different that ranking was from his "real" ones.

But yeah I'm not sure what you're referring to, if not to that.

I was referring to the BCS computer as an example but seriously...Sagarin's rankings go off of what he determined to be the most important factors? Unless he is right 100% of the time, it's still based on bias. He may have better reasoning behind his bias. He makes margin of victory very important. Sometimes you just get a fluke game that could skew those numbers. I'm really not trying to argue I just don't feel any ranking system is completely free of bias...how many things in life have a perfect way...

 

[RedDenver]

I think you could argue that computer rankings may be biased since they are designed by people. I mean why should strength of schedule be factored in but they pull out the margin of victory? Is it absolute fact that margin of victory means nothing? To the guy/group that wrote the program it is...

What? Are you referring to the BCS computers? Because those are obviously horrible. Any decent model whose goal is to predict results should include margin of victory. The fact that the BCS took margin of victory out of the equation was embarrassing.

I don't look at sh**ty models like that. Sagarin was a BCS computer, but he had to modify his formula and make an entirely new ranking, separate from the ones he posts each week on USA today, to give to the BCS. His normal rankings include margin of victory, and actually the majority of it is determined by margin of victory. But he had to take it out of the equation to appease the BCS. And it was hilarious how bad and how different that ranking was from his "real" ones.

But yeah I'm not sure what you're referring to, if not to that.

All predictive models have biases. Even the very best computer and statistical models are based on a human expert identifying what is and is not important (i.e. which stats get included and which do not). Even thinking that past performance includes some measure of future results is necessarily correct. I think you know know this - I'm just making the point that fancy math and algorithms doesn't change the inherent uncertainty in predictions or even the valuation of the data.

 

[tschu]

Sure. Any computer ranking just spits out what you tell it to do. But Sagarin's goal (as is the goal with any model) is to predict how many points better one team is than another and so on. And historically it has had remarkable correlation with the betting market which is the best measure of accuracy. Same with the F+ derived model. That's why I tend to pay attention to them much more than any other.

The larger point however is that a model treats all teams the same. It doesn't know the name of the team or any of their players or how that team did last year or ten years ago or whether it has 10 national championships or zero. It doesn't even know what conference that team is from. The team has a number and it has data points, and that's all there is to it.

And yes, like I posted before, there is uncertainty. Especially this early in the season, but even as time goes on there will be uncertainty. Another thing to remember - variance is a hell of a thing. And as a result, the better team does not win every game (sports would be boring as hell if they did) and scores can vary wildly. A model could 100% correctly have Team A favored by 7 over Team B, but Team B could win by 10 and the model wouldn't be wrong.

At the end of the day, sorry, I'm going to trust Sagarin over rando dude arguing that MSU sucks on the internet.