Sagarin

[UNH72Plus]

Well with season almost over and a couple of weeks before the finals, how about a little discussion about Sagarin and his latest rankings. Before I go any further, let state right up front that I greatly respect the football program at NDSU and believe that at this point in time they are best FCS program in the country. That being said, does anyone really believe that they are the 33rd best program in all of Division I? Could they really beat Notre Dame, Miami, and Virginia Tech? Perhaps, but I doubt it. And if you look his top 25 in the FCS, there's some real head scratchers. He ranks UNH (#82 overall/7th in FCS) 11 places below Chattanooga (#71/5), the team we beat in the quarter finals, and behind UNI (#58/3), SDSU (#68/4), and Coastal Carolina (#73/6). I'm sure I'll hear a lot about the strength of the MVFC, but UNI (#58) lost their last game with ISUr (#46/2) by 20 points while UNH only lost by3. Sagarin has UNI ranked 58 spots above Hawaii, a team that beat them. 9-5 South Dakota State got beat by 35 points by ISUr and their also ranked ahead of UNH. I guess the most shocking discrepancy though is his ranking of 4-8 Missouri State (#115/21) and 5-7 Western Illinois (#122/22) ahead of 9-5 Montana (#128/23) and 9-5 Richmond (#129/24). Missouri State and Western Illinois at 21st and 22nd in the FCS, really. You could also question 6-6 Southern Illinois being ranked 17th in the FCS. We hear a lot about the East Coast bypass on this site, but this looks a lot like a Midwest bypass to me. Again, let me state that I'm not trying to show any disrespect for NDSU or for that matter, any of the MVFC teams, but I really think you have to question where Sagarin comes up with these numbers.

Any thoughts?

[Milktruck74]

Bias or Bypass????

[ElCid]

Well with season almost over and a couple of weeks before the finals, how about a little discussion about Sagarin and his latest rankings. Before I go any further, let state right up front that I greatly respect the football program at NDSU and believe that at this point in time they are best FCS program in the country. That being said, does anyone really believe that they are the 33rd best program in all of Division I? Could they really beat Notre Dame, Miami, and Virginia Tech? Perhaps, but I doubt it. And if you look his top 25 in the FCS, there's some real head scratchers. He ranks UNH (#82 overall/7th in FCS) 11 places below Chattanooga (#71/5), the team we beat in the quarter finals, and behind UNI (#58/3), SDSU (#68/4), and Coastal Carolina (#73/6). I'm sure I'll hear a lot about the strength of the MVFC, but UNI (#58) lost their last game with ISUr (#46/2) by 20 points while UNH only lost by3. Sagarin has UNI ranked 58 spots above Hawaii, a team that beat them. 9-5 South Dakota State got beat by 35 points by ISUr and their also ranked ahead of UNH. I guess the most shocking discrepancy though is his ranking of 4-8 Missouri State (#115/21) and 5-7 Western Illinois (#122/22) ahead of 9-5 Montana (#128/23) and 9-5 Richmond (#129/24). Missouri State and Western Illinois at 21st and 22nd in the FCS, really. You could also question 6-6 Southern Illinois being ranked 17th in the FCS. We hear a lot about the East Coast bypass on this site, but this looks a lot like a Midwest bypass to me. Again, let me state that I'm not trying to show any disrespect for NDSU or for that matter, any of the MVFC teams, but I really think you have to question where Sagarin comes up with these numbers.

Any thoughts?

I am not a computer ranking defender, but have always used it as a data point. Above was one game out of a season. A game that was played at NH. Play it in Tennessee and it might have gone differently. It almost went differently in NH. I don't think that the fact that NH won invalidates the whole computer model. It has Chatty at 66.04 and NH at 63.64. That is only 2.5 points difference. In the case of UNI and Hawaii, same applies, play in Iowa and I think it would have been much different. The overall problem is that while all teams in the ranking are tied together by various degrees of separation, there are a couple teams which had very good games and they tend to skew the ratings. IN the case of UNI, the NDSU game figured big time. NDSU's position as a whole has lifted the MVFC higher. The other problem is that this system does not take into account lower Division games. These are sort of good examples of questionable team ranking, but I think you missed the biggest of all--Harvard???

[UNH Fanboi]

The problem with Sagarin and all computer rankings is that the sample size of OOC FCS games and FBS vs. FCS games is really small. Sagarin is useful as a general guideline, but should not be read as the end-all and be-all of determining which team is better than another.

[NY Crusader 2010]

This is a computer model that takes into account margin of victory. NDSU blew the doors off a Big 12 team by three touchdowns. Shouldn't be a surprise where they're ranked. Also remember that you add points for the home team. So NDSU would be predicted to lose to Notre Dame, Miami, Boise State and VT assuming they played on the road (which they would). They lifted the rest of the MVFC up and rightfully so. It's damn good league.

[NY Crusader 2010]

I think you missed the biggest of all--Harvard???

Why does everyone think Harvard is so bad? Didn't they go UNDEFEATED? They're ranked 105th, just a sniff above Samford, Liberty, SIU and Youngstown. People on this site seem to think the Ivy League is the same product as the Pioneer League and old MAAC which simply isn't the case. The third place team in this league beat an FBS team. All the Harvard alum in the NFL must be fakers who just fudged their resume and really went to school in the SoCon.

And this isn't coming from a Patriot-Ivy schill in denial. You know who I think is overrated by Sagarin? FORDHAM. They beat nobody this year and ran up big numbers on bad teams. They were tested for real three times and lost by double digits each time. Patriot League stunk this year, otherwise we wouldn't have even thought of the Rams as a top 10 team.

[FormerPokeCenter]

I agree with Sagarin's rankings, generally, and with regard to NDSU specifically....

McNeese took a 9-3 Nebraska squad to the wire, but finished 6-5 and didn't make it out of the Southland Conference, losing pretty badly to Sam Houston, who lost to NDSU.

And, let's be real. It's not like NDSU's peformances against FBS teams has been a fluke. It's been pretty consistent, so, yeah, I'd rank them in the top 40 in the country, regardless of division.

Easily...

[centennial]

Just want to say I used mostly Sagarin and Massey to pick in the FCS playoff challenge. Every now and then I will take chance on a team but I am seldomly able to beat the computers. I am 4th out of 120 this time and I think I was to 10 last year too. This year the only surprise was SHSU. I also picked Indiana state to beat chatty. Just that one pick would have gotten me close to number 1. Simple math dictates that I had a about 1 out of 300 chance of doing this.

[ElCid]

Why does everyone think Harvard is so bad? Didn't they go UNDEFEATED? They're ranked 105th, just a sniff above Samford, Liberty, SIU and Youngstown. People on this site seem to think the Ivy League is the same product as the Pioneer League and old MAAC which simply isn't the case. The third place team in this league beat an FBS team. All the Harvard alum in the NFL must be fakers who just fudged their resume and really went to school in the SoCon.

And this isn't coming from a Patriot-Ivy schill in denial. You know who I think is overrated by Sagarin? FORDHAM. They beat nobody this year and ran up big numbers on bad teams. They were tested for real three times and lost by double digits each time. Patriot League stunk this year, otherwise we wouldn't have even thought of the Rams as a top 10 team.

I think the same could be said of Harvard (sag schedule strength of 227/252 Div I...or in other words, opponents rated just about the upper end of the Pioneer League ratings). Who did Harvard beat to justify the ratings/rankings they have? Nobody. And the FBS team Yale beat was Army. A good win, but not exactly a high quality win, especially at home. I think folks would give the Ivy more respect if they stepped up and played a few worth wild OOC teams. Yale did with Army, and last year with Cal Poly (and got the win), but that is about it. The Ivy, in regard to computer rankings, has benefited from being insulated by their lack of a consistently meaningful OOC schedule. I still can't figure out how Harvard achieved a rating at 105. Nothing in this year's schedule should justify that.

But I agree, in hindsight, Fordam has been overrated this year. But then again, in regard to polls, there are many teams that live off of reputation rather than performance.

[Bisonwinagn]

I think these rankings along with the GPI seem pretty accurate when looking at teams. I think NDSU was 17th last year if I remember correctly and would have competed with teams outside of the top 10 in a one game set. It's the same argument Boise has had in the BCS rankings.

[CasualFan]

I believe Sagarin uses some sort of least squares model. So it will try to put you below teams you lost to, and above teams you beat, but has to tolerate some error. Think of games between three teams, where each one loses to one opponent and beats another. The model probably looks for other things to decide which one of the three goes on top. With these types of models, you can always find examples of inversion (ranked lower than a team they beat) due to the minimizing of the error in placement. Just because a team is ranked higher, doesn't mean they always will beat a team ranked lower. Any Given Saturday, right?

I'm just guessing, but I bet Harvard's opponents are so low that even though they beat them all, they just have to be placed above the highest opponent. Coastal Carolina had the same problem. I believe CCU's rank increased during their last three games including the playoffs, even though they lost 2 of 3, because they were playing more difficult competition and their losses were close.

[Hammersmith]

I think the biggest problem with ranking Harvard is the limited non-conference Ivy slate, and the incestuous nature of their relationship with the Patriot. There just aren't enough data points to accurately rank the Ivies. The Ivies averaged just over 1 OOC game each outside the PL(8 Ivy teams, 10 non-PL/Ivy games). When it comes down to it, that's about the same or worse as the amount of FBS to FCS linkage. I haven't counted out the games, but I think the SWAC & MEAC are in a similar bubble.


And while I like to look at the GPI(and sometimes I create my 'GPI with 1/3 less spite' version), I have to laugh at Ralph bragging it up in the other thread. Sagarin is original work. Massey(not the composite) is original work. All the polls are original work. All the other computer rankings are original work. GPI is not. GPI is just a slightly modified average of other peoples' work. If the GPI accurately predicts something, it's because other people did the work. It's still a good tool to look at(especially if you add the AGS poll back in), but Ralph is trying to make everybody believe it's far more original than it actually is.

/minirant

[NY Crusader 2010]

Hammersmith it is true that the Ivy keeps itself in a scheduling bubble but they do play a lot of games against the PL who does play a fair share of games against other FCS teams. So there is at least some level of linkage by the second degree. Couldn't you say the same about Big Sky teams playing tons of NAIA and DIIs do to regional necessity?

[semobison]

Well with season almost over and a couple of weeks before the finals, how about a little discussion about Sagarin and his latest rankings. Before I go any further, let state right up front that I greatly respect the football program at NDSU and believe that at this point in time they are best FCS program in the country. That being said, does anyone really believe that they are the 33rd best program in all of Division I? Could they really beat Notre Dame, Miami, and Virginia Tech? Perhaps, but I doubt it. And if you look his top 25 in the FCS, there's some real head scratchers. He ranks UNH (#82 overall/7th in FCS) 11 places below Chattanooga (#71/5), the team we beat in the quarter finals, and behind UNI (#58/3), SDSU (#68/4), and Coastal Carolina (#73/6). I'm sure I'll hear a lot about the strength of the MVFC, but UNI (#58) lost their last game with ISUr (#46/2) by 20 points while UNH only lost by3. Sagarin has UNI ranked 58 spots above Hawaii, a team that beat them. 9-5 South Dakota State got beat by 35 points by ISUr and their also ranked ahead of UNH. I guess the most shocking discrepancy though is his ranking of 4-8 Missouri State (#115/21) and 5-7 Western Illinois (#122/22) ahead of 9-5 Montana (#128/23) and 9-5 Richmond (#129/24). Missouri State and Western Illinois at 21st and 22nd in the FCS, really. You could also question 6-6 Southern Illinois being ranked 17th in the FCS. We hear a lot about the East Coast bypass on this site, but this looks a lot like a Midwest bypass to me. Again, let me state that I'm not trying to show any disrespect for NDSU or for that matter, any of the MVFC teams, but I really think you have to question where Sagarin comes up with these numbers.

Any thoughts?

Don't kid yourself, WIU And Missouri State were very tough football teams this year. WIU lost to two Big Ten teams, trailed Wisconsin 9-3 at the half holding Melvin Gordon to 51 yards on 17 carries. They also led NDSU 10-3 after 3 quarters. Missouri State lost to Oklahoma State 40-23 and lost a lot of close games in MVFC play early. South Dakota was the only easy game for most teams in Valley play this year!

[NY Crusader 2010]

Don't kid yourself, WIU And Missouri State were very tough football teams this year. WIU lost to two Big Ten teams, trailed Wisconsin 9-3 at the half holding Melvin Gordon to 51 yards on 17 carries. They also led NDSU 10-3 after 3 quarters. Missouri State lost to Oklahoma State 40-23 and lost a lot of close games in MVFC play early. South Dakota was the only easy game for most teams in Valley play this year!

Agreed. This year's version of the MVFC might be the deepest and best conference in I-AA / FCS history.

[caribbeanhen]

Agreed. This year's version of the MVFC might be the deepest and best conference in I-AA / FCS history.

that's a nice Christmas present you just presented to the MVFC........ A pure gift

[underdawg]

And well deserved I might add

[Wallace]

And while I like to look at the GPI(and sometimes I create my 'GPI with 1/3 less spite' version), I have to laugh at Ralph bragging it up in the other thread. Sagarin is original work. Massey(not the composite) is original work. All the polls are original work. All the other computer rankings are original work. GPI is not. GPI is just a slightly modified average of other peoples' work. If the GPI accurately predicts something, it's because other people did the work. It's still a good tool to look at(especially if you add the AGS poll back in), but Ralph is trying to make everybody believe it's far more original than it actually is./minirant
I didn't brag or try to make anyone "believe" anything. The GPI is an index, it is stated in the name. It is the first one specifically for the FCS. It is not a predictor, it indicates what the specifically chosen contents of the index say about team rankings.

[skinny_uncle]

The GPI is a composite of selected other rankings. These are often more accurate than individual rankings because they average out some of the biases. The idea probably originated in the Massey composite. The GPI just uses a smaller sample.