Sagarin Ratings for BCS and FCS

[AlphaSigMD]

Just thought I'd post up the lastest computer wizardry by JS.

ASU is nearly 40 positions ahead of every other FCS school (closest - NDSU, congrats), and higher ranked than all but 37 BCS teams (for now).

Its a given that ASU will drop some because of SOS, but its nice seeing us there in the thick of things.

I also predict michigan will go back up into the top 20 when they trounce oregon later this week.

As a side note, among the 37 teams ranked above ASU in the ratings, I believe that ASU would have a feasible shot of taking out the following teams.

1. VaTech (ASU is a few points better than ECU)
2. Texas (and several points better than Ark St.)
3. Purdue (Big 10 has a history of being ineffective)
4. BYU (Hit and miss at best, in the Mountain West)
5. Washington (The PAC 10 has no defense and suspect running game)

Yes, I know...you all think I'm crazy.

[McNeese75]

Just thought I'd post up the lastest computer wizardry by JS.

ASU is nearly 40 positions ahead of every other FCS school (closest - NDSU, congrats), and higher ranked than all but 37 BCS teams (for now).

Its a given that ASU will drop some because of SOS, but its nice seeing us there in the thick of things.

I also predict michigan will go back up into the top 20 when they trounce oregon later this week.

As a side note, among the 37 teams ranked above ASU in the ratings, I believe that ASU would have a feasible shot of taking out the following teams.

1. VaTech (ASU is a few points better than ECU)
2. Texas (and several points better than Ark St.)
3. Purdue (Big 10 has a history of being ineffective)
4. BYU (Hit and miss at best, in the Mountain West)
5. Washington (The PAC 10 has no defense and suspect running game)

Yes, I know...you all think I'm crazy.

I think it would be a whole different situation if ASU were to play any of those teams at this point in time. You can bet none of them would be looking down at ASU and they would be ready and pouring the coals to them from the opening kickoff xnodx

[Purple Knight]

According to Saragin, Furman is 44th FCS team in nation. Don't think I would get too pumped up with his system.

[AZGrizFan]

Just thought I'd post up the lastest computer wizardry by JS.

ASU is nearly 40 positions ahead of every other FCS school (closest - NDSU, congrats), and higher ranked than all but 37 BCS teams (for now).

Its a given that ASU will drop some because of SOS, but its nice seeing us there in the thick of things.

I also predict michigan will go back up into the top 20 when they trounce oregon later this week.

As a side note, among the 37 teams ranked above ASU in the ratings, I believe that ASU would have a feasible shot of taking out the following teams.

1. VaTech (ASU is a few points better than ECU)
2. Texas (and several points better than Ark St.)
3. Purdue (Big 10 has a history of being ineffective)
4. BYU (Hit and miss at best, in the Mountain West)
5. Washington (The PAC 10 has no defense and suspect running game)

Yes, I know...you all think I'm crazy.

And I'm gonna go out on a limb and say that if ANY team on the list above you was NEXT on your schedule, you wouldn't beat a single one of them. You ain't surprising anybody now, Doc. xeyebrowx ;)

[JohnStOnge]

And I'm gonna go out on a limb and say that if ANY team on the list above you was NEXT on your schedule, you wouldn't beat a single one of them. You ain't surprising anybody now, Doc. xeyebrowx ;)

I don't know about that. If I had to bet straight up I'd bet on any of those listed but I sure wouldn't be absolutely shocked if App beat any of them (especially BYU) except Virginia Tech. I'd have said Texas if Texas hadn't been fortunate to beat Arkansas State.

Virginia Tech has handled the "West Virginia" (actually Tulane from the days of that staff there) style offense pretty well in the past. So I would be a bit more shocked if App could beat the Hokies under the current circumstances in which they just caught everybody's attention.

[gr8ness97]

link?

[JohnStOnge]

By the way, that #38 rating is not an unprecedented Sagarin rating for a I-AA/FCS. I can't document it because the web site doesn't go back that far but my recollection is that Marshall's 1996 I-AA team finished that season rated 21st by that system. The second highest I can recall is James Madison's 2004 team...which ended at 36th.

Delaware's 2003 team ended at 40th.

You can see those by going to Sagarin's site and clicking on "Archives."

[KAUMASS]

Not taking any credit away from App. State, those rankings will have more bearing 1/2 through the year as more teams play each other, and the rankings will settle down. Kind of like batting averages for major league players who get 500 at bats. A hit early in the year can jump their average 20-50 points, but towards the end it can up their average about 1 point.

Those power rankings will be more accurate as the season progresses.

[youwouldno]

There's no mystery to Sagarin ratings. It's just a regression line drawn through plot points, which are of course football games. When the number of data points is low, the regression line is wildly inaccurate and Sagarin would say the same thing. So the 'problem,' such as it is, has less to do with the system and more to do with the people reading the results, while being unable to understand the basic math going on.

App St's rating could get pretty high if Michigan does well. If both teams run the table, App could finish pretty high, at which point Sagarin would overrate them (in the same way San Diego was overrated- no 'lower bound' data points).

[DuckDuckGriz]

Just thought I'd post up the lastest computer wizardry by JS.

ASU is nearly 40 positions ahead of every other FCS school (closest - NDSU, congrats), and higher ranked than all but 37 BCS teams (for now).

Its a given that ASU will drop some because of SOS, but its nice seeing us there in the thick of things.

I also predict michigan will go back up into the top 20 when they trounce oregon later this week.

As a side note, among the 37 teams ranked above ASU in the ratings, I believe that ASU would have a feasible shot of taking out the following teams.

1. VaTech (ASU is a few points better than ECU)
2. Texas (and several points better than Ark St.)
3. Purdue (Big 10 has a history of being ineffective)
4. BYU (Hit and miss at best, in the Mountain West)
5. Washington (The PAC 10 has no defense and suspect running game)

Yes, I know...you all think I'm crazy.

Not to rain on the parade, but I highly doubt all of this.

ASU's offensive style may have surprised Big Ten Michigan but it's nothing Pac-10 defenses don't see week after week.

[JohnStOnge]

Not taking any credit away from App. State, those rankings will have more bearing 1/2 through the year as more teams play each other, and the rankings will settle down. Kind of like batting averages for major league players who get 500 at bats. A hit early in the year can jump their average 20-50 points, but towards the end it can up their average about 1 point.

Those power rankings will be more accurate as the season progresses.

Not that much more (I've studied that). But right now Sagarin's system is not totally objective. When he says he's using "bayesian," that means he's using information other than objective assessment of what teams have accomplished this season (pretty obvious, since he had ratings before the season even started).

[EmeryZach]

The Sagarin poll only makes sense when the season is over

[OL FU]

There's no mystery to Sagarin ratings. It's just a regression line drawn through plot points, which are of course football games. When the number of data points is low, the regression line is wildly inaccurate and Sagarin would say the same thing. So the 'problem,' such as it is, has less to do with the system and more to do with the people reading the results, while being unable to understand the basic math going on.
App St's rating could get pretty high if Michigan does well. If both teams run the table, App could finish pretty high, at which point Sagarin would overrate them (in the same way San Diego was overrated- no 'lower bound' data points).

I agree with everything you said except the bold. The problem is not the people reading the results. Statistics are used to draw conclusions. Sagarin list the teams in order of their power ranking therefore the conclusion one would draw is that, at this point in time, Sagarin thinks this is the ranking of each college team. If he agrees the numbers at present are meaningless he should put in big bold letters :


This ranking means nothing until week ???? whatever the week is

[HensRock]

He got one thing right:

38. Appalachian State
40. Michigan

[OL FU]

He got one thing right:

38. Appalachian State
40. Michigan

but after the fact just like the rest of usxsmiley_wix

[FormerPokeCenter]

Actually, there IS a disclaimer in the Sagarin Rankings that the ranking system isn't fully indexed until week four, I believe, when there are enough common opponents for the rankings to have relevance. The rankings after week one are notoriously skewed.....

[OL FU]

Actually, there IS a disclaimer in the Sagarin Rankings that the ranking system isn't fully indexed until week four, I believe, when there are enough common opponents for the rankings to have relevance. The rankings after week one are notoriously skewed.....

Good for himxthumbsupx

[OL FU]

Actually, there IS a disclaimer in the Sagarin Rankings that the ranking system isn't fully indexed until week four, I believe, when there are enough common opponents for the rankings to have relevance. The rankings after week one are notoriously skewed.....

This is all I could find. Not exactly a disclaimer unless I missed it


For the first few weeks of the season, the starting ratings have weight
in the process(BAYESIAN), but once the teams are all CONNECTED, then
the starting ratings are no longer used and all teams are started equal
and the ELO-CHESS is then done in an UNBIASED manner from that point on.
RATING and PREDICTOR are now bayesian and the ELO-CHESS is also bayesian

[FormerPokeCenter]

Somewhere back in the recesses of my memory, I seem to recall that week 4 was the magic week for interconnectedness in the Sagarin. Whether or not that came from a Jeff Sagarin quote somewhere, or what, I do not recall...

JohnStOnge and I had a similar discussion about the Sagarin several years go on GeauxCowboys.com. I think his memory's probably better than mine and, plus, he's a statistical analysis kinda guy, whereas I have difficulty with all things numeric, including balancing my checkbook....

I think if you look in the initial text of the Sagarin, there are references to the various style of analyses used being "BIASED" and I think in week 4, most - if not all - show up as "are now UNBIASED."

[FormerPokeCenter]

I spoke out of school. The different types of analyses are refered to as "Bayesian" rather than biased for the first few weeks of the season. After week 4, they're "Unbiased."

What does this "bayesian" reference mean?

This, from Wikipedia:

"Bayesian probability is an interpretation of the probability calculus which holds that the concept of probability can be defined as the degree to which a person (or community) believes that a proposition is true. Bayesian theory also suggests that Bayes' theorem can be used as a rule to infer or update the degree of belief in light of new information"

I think that whole reference to "believes that a proposition is true" could reasponably be interpreted as "bias."

So, rather than saying, "I'm flying by the seat of my pants, using conventional wisdom, based on what most football fans believe to be the truth" for the first 4 weeks of the season, he's substituting "Bayesian probability analysis"....

[bisonbill]

If Appy played Michigan 10 more times this season.... Michigan would win every game.

[Gil Dobie]

If Appy played Michigan 10 more times this season.... Michigan would win every game.

Would Michigan get a home game everytime? Nice weather everytime?

The game was played and App St won.............PERIOD. xthumbsupx

[AlphaSigMD]

If Appy played Michigan 10 more times this season.... Michigan would win every game.

Saying this statement is just as ludicrous and unfounded as saying "If App played NDSU 10 times this season....App State would win every game."

We will never know, because neither situation will occur.

However, I'd love for ASU to play NDSU anytime, anywhere.

[appfan2008]

id love to get a asu ndsu game on the schedule...

[X-Factor]

Saying this statement is just as ludicrous and unfounded as saying "If App played NDSU 10 times this season....App State would win every game."

We will never know, because neither situation will occur.

However, I'd love for ASU to play NDSU anytime, anywhere.

You guys had your chance...we were VERY generous money wise with our offer to play Sept 1 at the Fargodome but you guys decided to play Michigan instead. Would have been one heck of a game. Can't say it was a bad decision for you guys to turn it down after the fact though. :D

[skinny_uncle]

If Appy played Michigan 10 more times this season.... Michigan would win every game.
I haven't seen anyone from Michigan demanding a rematch.
xeyebrowx

[bisonbill]

If Appy played Michigan 10 more times this season.... Michigan would win every game.

Appy State just like every other very good FCS team given the right situation can beat a ranked FBS team 1 out of 30 times. What I am saying is that they played that 1 game first.

[Gil Dobie]

Appy State just like every other very good FCS team given the right situation can beat a ranked FBS team 1 out of 30 times. What I am saying is that they played that 1 game first.

If 30 games were possible, I'd put a lot of money on App St winning a 2nd game. App St is a great FCS team, not just good. Michigan isn't that much better than App St.

[bisonbill]

If 30 games were possible, I'd put a lot of money on App St winning a 2nd game. App St is a great FCS team, not just good. Michigan isn't that much better than App St.

How many times have FCS teams played ranked FBS teams over the years? How many games have been won by the FCS team? The odds are probably closer to 100-1 if you look at the history. I agree if it was possible to play 100 games between the CURRENT Appy State and Michigan teams that Appy state would win a lot more than one. My point is that the sun and moon have to align for FCS teams to beat ranked FBS teams. It is obvious since it has happened only once.

[SirApp]

How many times have FCS teams played ranked FBS teams over the years? How many games have been won by the FCS team? The odds are probably closer to 100-1 if you look at the history. I agree if it was possible to play 100 games between the CURRENT Appy State and Michigan teams that Appy state would win a lot more than one. My point is that the sun and moon have to align for FCS teams to beat ranked FBS teams. It is obvious since it has happened only once.

Didn't realize that there was an eclipse on Saturday....

[bisonbill]

Didn't realize that there was an eclipse on Saturday....

I did not indicate which solar system the sun and moon were from.

[Gil Dobie]

How many times have FCS teams played ranked FBS teams over the years? How many games have been won by the FCS team? The odds are probably closer to 100-1 if you look at the history. I agree if it was possible to play 100 games between the CURRENT Appy State and Michigan teams that Appy state would win a lot more than one. My point is that the sun and moon have to align for FCS teams to beat ranked FBS teams. It is obvious since it has happened only once.

How many times has a team as great as App St played a ranked FCS school? Not much history there. I trying not to belittle App St's win, but Michigan had trouble with Ball St, and gave up a lot of points in their last 3 games. That hints to me that Michigan is probably not a top 5 team. With their recruiting numbers, and the money they spend on football, they should be. They'll be lucky to finish 3 or 4 in the Big Ten behind Ohio St, Wisconsin, and maybe Penn St, or another team that moves up like Purdue or Illinois. Great win by App St over a traditional College football power.

[CopperCat]

Just thought I'd post up the lastest computer wizardry by JS.

ASU is nearly 40 positions ahead of every other FCS school (closest - NDSU, congrats), and higher ranked than all but 37 BCS teams (for now).

Its a given that ASU will drop some because of SOS, but its nice seeing us there in the thick of things.

I also predict michigan will go back up into the top 20 when they trounce oregon later this week.

As a side note, among the 37 teams ranked above ASU in the ratings, I believe that ASU would have a feasible shot of taking out the following teams.

1. VaTech (ASU is a few points better than ECU)
2. Texas (and several points better than Ark St.)
3. Purdue (Big 10 has a history of being ineffective)
4. BYU (Hit and miss at best, in the Mountain West)
5. Washington (The PAC 10 has no defense and suspect running game)

Yes, I know...you all think I'm crazy.

Proof that Sagarin ratings don't mean a damn thing.

You can speculate all you want at who App would beat among those teams. But you ain't gonna beat them. Just because you beat Michigan doesn't mean that you can beat them all.