SP+ Connelly

Donate To GopherHole
S

swingman

Well-known member
Joined
Aug 12, 2010
Messages
3,143
Reaction score
2,022
Points
113
SP loves NE, WI, MD...Gophs back in top 40...would be slight underdogs to NE and MD at Bank..

From Connelly: Four top-10 teams -- and nine ranked teams in all -- lost in week five of the 2021 college football season, and the wackiness and chaos levels are higher than they've been in years. As you would expect, week five's unpredictability has resulted in a lot of movement in the SP+ ratings.

SP+ 2021 History: Preseason | Week 1 | Week 2 | Week 3 | Week 4

Click here for the full rankings after Week 5.

Overall SP+ Rankings
TEAM
RATING OFFENSE DEFENSE SPECIAL TEAMS

3. Ohio St. (4-1) 26.3 46.7 (1) 21.0 (27) 0.6 (2)
6. Penn St. (5-0) 21.6 35.3 (23) 13.8 (5) 0.1 (54)
8. Michigan (5-0) 19.7 34.4 (26) 15.3 (8) 0.6 (3)
10. Wisconsin (1-3) 17.6 29.2 (63) 11.7 (3) 0.2 (32)
11. Cincinnati (4-0) 17.2 33.7 (33) 16.1 (9) -0.4 (123)
12. Iowa St. (3-2) 17.0 34.3 (27) 17.2 (13) 0.0 (70)
13. Pittsburgh (4-1) 16.7 37.4 (11) 21.1 (29) 0.3 (14)
15. Iowa (5-0) 15.4 27.6 (72) 12.6 (4) 0.4 (8)
17. Texas A&M (3-2) 14.7 30.7 (51) 16.6 (11) 0.5 (5)
20. Notre Dame (4-1) 14.0 32.2 (41) 18.3 (17) 0.0 (58)
21. Maryland (4-1) 13.9 35.9 (18) 21.8 (33) -0.2 (102)
22. Nebraska (3-3) 13.5 33.3 (34) 19.4 (20) -0.4 (124)
24. Michigan St. (5-0) 13.3 33.0 (38) 19.7 (24) 0.0 (64)
25. Miami (2-3) 13.1 35.3 (24) 22.2 (37) 0.0 (59)
26. Liberty (4-1) 12.9 35.9 (17) 22.8 (41) -0.2 (94)
39. Minnesota (3-2) 9.5 32.2 (40) 22.9 (42) 0.1 (48)
54. Purdue (3-2) 6.6 29.2 (62) 22.6 (39) -0.1 (72)
59. Indiana (2-3) 5.4 27.5 (73) 22.0 (35) -0.1 (84)
62. Rutgers (3-2) 5.1 26.7 (82) 21.4 (31) -0.1 (93)
86. Northwestern (2-3) -1.9 21.5 (107) 23.2 (47) -0.2 (103)
91. Miami-OH (2-3) -4.3 25.8 (87) 30.2 (93) 0.1 (37)
92. Illinois (2-4) -4.9 24.0 (100) 29.3 (88) 0.4 (12)
105. Colorado (1-4) -9.3 20.4 (116) 29.6 (90) -0.1 (85)
113. BGSU (2-3) -11.4 14.4 (129) 26.0 (64) 0.3 (17)
 

From Connelly:

Biggest movers

As we work our way into October, conference-level adjustments -- in which I look at how much initial SP+ ratings are missing on average for teams from given conferences and adjust accordingly (kind of like how house prices are determined in part by the neighborhood in which you live) -- begin to kick in a lot stronger. What does that mean for Week 5? Very good things for the Big Ten, which saw its teams receive healthy adjustments thanks to quality performance against projections.

Three Big Ten ratings are perhaps most noteworthy.

No. 3 Ohio State. The Buckeyes have beaten their last two opponents by a combined 111-20 and seem to be finding their defensive rhythm. Granted, walloping Akron and Rutgers doesn't tell us everything we need to know, but it was impressive, and combined with the Big Ten's friendly adjustments, Ohio State damn near caught Alabama for No. 2.

No. 10 Wisconsin. SP+ just can't quit the Badgers, who have now lost two games to higher-ranked Big Ten teams (Penn State by a small margin, Michigan by a larger one) and suffered a fluky, return-score-addled loss to Notre Dame. It isn't sold on their faulty offense, obviously, but the defense has been so good that, again with help from the Big Ten adjustment, they remain more highly-ranked than the eyeballs suggest they should.

No. 22 Nebraska. SP+ remained higher on the Huskers than most despite their 2-3 start, but they charged into the top 25 following their 56-7 humiliation of Northwestern on Saturday. They have obvious issues in close games, and have for most of Scott Frost's tenure, but play for play and drive for drive, they have been awfully good this year.

Biggest rises

Twelve teams rose by at least 10 spots this week.

Nebraska: up 25 spots from 47th to 22nd

Fresno State: up 18 spots from 59th to 41st

Minnesota: up 17 spots from 56th to 39th

Iowa: up 13 spots from 28th to 15th

Tennessee: up 13 spots from 32nd to 19th

South Carolina: up 13 spots from 86th to 73rd

FAU: up 13 spots from 88th to 75th

Virginia: up 11 spots from 58th to 47th

Florida State: up 10 spots from 73rd to 63rd

Houston: up 10 spots from 75th to 65th

Air Force: up 10 spots from 76th to 66th

Ball State: up 10 spots from 118th to 108th

The top of this list has a Big Ten flavor for the reasons mentioned above, but Fresno State jumped up despite a loss to Hawaii late on Saturday night. Why? First, with each game you play, your preseason projections carry less weight in the SP+ formula, so the Bulldogs' impressive early-season performances carried more weight this week and boosted them a bit. Second, SP+ viewed the loss to Hawaii as a pretty fluky occurrence. The Bulldogs outgained the Rainbow Warriors by 157 yards and 1.9 yards per play but suffered an unsustainable minus-5 turnover margin. It still trusts Fresno a good amount.

Biggest stumbles

Thirteen teams dropped by at least 11 spots

UCLA: down 22 spots from 10th to 32nd

SMU: down 15 spots from 40th to 55th

Missouri: down 15 spots from 52nd to 67th

Army: down 15 spots from 64th to 79th

Oregon: down 14 spots from 17th to 31st

Oregon State: down 14 spots from 43rd to 57th

California: down 14 spots from 67th to 81st

UAB: down 13 spots from 55th to 68th

Troy: down 13 spots from 74th to 87th

Arkansas: down 12 spots from 16th to 28th

Washington: down 11 spots from 38th to 49th

Tulane: down 11 spots from 77th to 88th

Colorado: down 11 spots from 94th to 105th

The presence of six Pac-12 teams on this list should pretty definitively tell you which conference suffered the most from the increased weight of conference-level adjustments.
 

I’d rather be 3-2 and ranked lower in whatever this is than be 1-3 and ranked 10th
 







Don't remember who posted the, "This is a bad win..." post, and it looks like it has since been deleted, but this is what he was talking about. I was pretty surprised that the Gophers moved up after Purdue, but think that says more about what happened against BGSU.

SP+ shouldn't have nice things to say about your odds of winning the next game when you get outgained by 100 yards, complete 50% of your passes, and rush for less than 4 YPC. Nebraska looked very good this week, and think they'll be favorites at MN in two weeks.
 



Bob_Loblaw said:
Whatever formula they are using needs to be revisited.

Wisconsin (1-3) is ranked higher than Cinci (4-0).
Click to expand...
It's not a ranking. It's a forward-looking projection of who would win if they played on a neutral field.

If you take the first number after each teams record (e.g., 9.5 for the Gophers), and compare it to another team (e.g., 13.5 for Nebraska), then the difference should be the neutral field spread (for example, the Gophers should be 4-point dogs on a neutral field, and 1-point dogs at home).

This formula is accurate enough to beat the Vegas spread about 55% of the time.
 
Last edited:

hello-world said:
It's not a ranking. It's a forward-looking projection of who would win if they played on a neutral field.
Click to expand...
It is a ranking (teams are ordered from 1 to 100+ according to their rank). I get that the rank is forward looking. What I'm saying is that whatever formula they are using to put the teams into particular rankings (of who would win at a neutral site), is broken.

If Cinci and Wisconsin played on a neutral field next week, Cinci should be favored by a TD.
 


hello-world said:
Don't remember who posted the, "This is a bad win..." post, and it looks like it has since been deleted, but this is what he was talking about. I was pretty surprised that the Gophers moved up after Purdue, but think that says more about what happened against BGSU.

SP+ shouldn't have nice things to say about your odds of winning the next game when you get outgained by 100 yards, complete 50% of your passes, and rush for less than 4 YPC. Nebraska looked very good this week, and think they'll be favorites at MN in two weeks.
Click to expand...

SP doesn't mind getting beat by 3 TDs, completing under 50% of your passes and rushing for under 2 yards though (and that's your best game in two weeks).
 



Bob_Loblaw said:
It is a ranking (teams are ordered from 1 to 100+ according to their rank). I get that the rank is forward looking. What I'm saying is that whatever formula they are using to put the teams into particular rankings (of who would win at a neutral site), is broken.

If Cinci and Wisconsin played on a neutral field next week, Cinci should be favored by a TD.
Click to expand...
Yeah idk man, his system works so I tend to trust it
Bob_Loblaw said:
SP doesn't mind getting beat by 3 TDs, completing under 50% of your passes and rushing for under 2 yards though (and that's your best game in two weeks)
Click to expand...
You're comparing a post I made about Minnesota to a completely unrelated game.

The difference is that Wisconsin was playing what SP+ believes is the 8th best defense in the country while Minnesota was playing the 40th.
 

Bob_Loblaw said:
Whatever formula they are using needs to be revisited.

Wisconsin (1-3) is ranked higher than Cinci (4-0).
Click to expand...
The problem with S&P is it’s pretty much useless the first 2/3rds of the year because it includes data that helped them give their preseason rankings. It also is pretty useless early in the year because SOS is such a factor…but the SOS also includes data from last season that informs preseason rankings.

im not sure which week they stop including previous years data…but that’s the week it STARTS to be interesting to look at.

Even still, trying to rank 130 teams based on 4-6 games each there just isn’t enough data. Not enough crossover.
 

hello-world said:
Yeah idk man, his system works so I tend to trust it

You're comparing a post I made about Minnesota to a completely unrelated game.

The difference is that Wisconsin was playing what SP+ believes is the 8th best defense in the country while Minnesota was playing the 40th.
Click to expand...
That's cool, I'm not asking anyone else not to trust it. It feels like the trust is a little blind. I cannot think of a statistic that could possibly point you to thinking Wisconsin is a better football team than Cinci except for things completely unrelated to their play on the field this year. This isn't just me hating on our rivals, I think Iowa is ranked too low.

I get that I'm comparing a completely unrelated game however we are talking about the formula used to rank all of the teams. You pointed to SP not looking kindly on the Gophers ugly win because of the various reasons you listed. You seem to be talking about how the formula will look at the Gophers (you were bringing variables that one would assume would be used in the calculation). However, they still have a team ranked in the TOP 10 that was exponentially worse in those same categories than the Gophers. I'm not arguing that strength of an opponent shouldn't play a factor, I'm saying that no team that has back-to-back weeks like the Badgers just had should be ranked in the top 10.

If SP had the Badgers higher than the Gophers, that would still make sense to me. They were embarrassed by pretty good (not great) football team. I get that wouldn't hurt them much worse than a road win. However, the fact that the Badgers are higher than Cinci and Iowa should make people second guess the formula a bit.
 

Some guy said:
im not sure which week they stop including previous years data…but that’s the week it STARTS to be interesting to look at.
Click to expand...

He stopped writing it out because it's more accurate when he includes it. Just gets less important.
 

Bob_Loblaw said:
I cannot think of a statistic that could possibly point you to thinking Wisconsin is a better football team than Cinci except for things completely unrelated to their play on the field this year.
Click to expand...
Yeah - I mean, he includes recruiting rankings, past season performance, etc. so that might be what's doing it.
 

hello-world said:
He stopped writing it out because it's more accurate when he includes it. Just gets less important.
Click to expand...
Any time you are using things that didn’t occur this season to rank teams this season, it’s a stupid ranking.


it is not stupid for predictive reasons.
it is stupid for ranking reasons.


Personally I don’t really care who should hypothetically win a game that won’t happen. I prefer to analyze what did happen. Same reason I don’t want 2 teams from the same conference in the playoff. Same reason I don’t want a 18-12 (10-10) big ten team in the ncaa tournament over a 27-3 team from a smaller conference.
 

Some guy said:
Any time you are using things that didn’t occur this season to rank teams this season, it’s a stupid ranking.


it is not stupid for predictive reasons.
it is stupid for ranking reasons.


Personally I don’t really care who should hypothetically win a game that won’t happen. I prefer to analyze what did happen. Same reason I don’t want 2 teams from the same conference in the playoff. Same reason I don’t want a 18-12 (10-10) big ten team in the ncaa tournament over a 27-3 team from a smaller conference.
Click to expand...
This is why you shouldn't think of it as a ranking. It's a predictive tool that's useful for sports gambling and general interest. It's never going to get used to decide who deserves to be in the playoffs for example.
 

hello-world said:
This is why you shouldn't think of it as a ranking. It's a predictive tool that's useful for sports gambling and general interest. It's never going to get used to decide who deserves to be in the playoffs for example.
Click to expand...
I don't know why people keep saying this with S&P, it is a ranking. It's a different kind of ranking than like the AP or Coaches Poll, but it is a ranking.
 

hello-world said:
This is why you shouldn't think of it as a ranking. It's a predictive tool that's useful for sports gambling and general interest. It's never going to get used to decide who deserves to be in the playoffs for example.
Click to expand...
Except metrics like it really are.
People cite kenpom in basketball.
People cite sagarin
People cite “football power index”
They originally used “game control” in the playoff committee but have walked back officially sanctioning it


Efficiency based metrics weighted towards strength of schedule double bias towards their own metric (because the bias is in the metric and baked into the SOS).
Mathematically they reward teams that play well rather than teams that win. Usually these two are correlated. But not always. And winning should be more important than playing well.
 

How could he rank Iowa at 15? People want me to believe this is an all time great B1G team.
 

Bob_Loblaw said:
I don't know why people keep saying this with S&P, it is a ranking. It's a different kind of ranking than like the AP or Coaches Poll, but it is a ranking.
Click to expand...
Yeah and people that drive the narrative look at it and other efficiency metrics like it
 

Bob_Loblaw said:
I don't know why people keep saying this with S&P, it is a ranking. It's a different kind of ranking than like the AP or Coaches Poll, but it is a ranking.
Click to expand...
Because he isn't sitting down and stack ranking every team based on how they've played, nor is he comparing records. In fact, wins and losses don't matter at all.

He's essentially running a bunch of averages and regressions to try to predict, on average, how many points a team should be expected to score and give up in any given week.

The algorithm doesn't care who won any given game. In fact, it recalculates win probabilities after the game that are completely seperate from the actual outcome.

As long as it is beating Vegas spread, it's not broken, it's doing exactly what it's supposed to do.
 

hello-world said:
Because he isn't sitting down and stack ranking every team based on how they've played, nor is he comparing records. In fact, wins and losses don't matter at all.

He's essentially running a bunch of averages and regressions to try to predict, on average, how many points a team should be expected to score and give up in any given week.

The algorithm doesn't care who won any given game. In fact, it recalculates win probabilities after the game that are completely seperate from the actual outcome.

As long as it is beating Vegas spread, it's not broken, it's doing exactly what it's supposed to do.
Click to expand...
Yeah. You are totally spot on here.
The playoff committee shouldn’t be allowed to look at metrics where the wins and losses don’t matter
 

Some guy said:
Yeah. You are totally spot on here.
The playoff committee shouldn’t be allowed to look at metrics where the wins and losses don’t matter
Click to expand...
Right, like I follow his numbers pretty closely, because I like to bet on sports. I'm not going to sit here and argue that Wisconsin deserves to go to a better bowl game than Cincy at the end of the season.

fwiw; it's still high on the Badgers because it sees turnovers as "luck." Mertz threw two pick 6's and two other interceptions in his own half. SP+ isn't going to punish the Badgers for that because it isn't likely to happen if they played again.

We see 41-13. SP+ sees an incredibly unlucky performance.
 
Last edited:

You must log in or register to reply here.



Top Bottom