B1G Ten Standings - Deeper Look

[caliGopher]

Below is the standing in the B1G as of yesterday's games. I keep track of a number of stats, because it's fun, and with the bigger conference, it helps put perspective since any teams record in the new B1G is not impacted by 8 teams.

What started this was the argument that Alabama is the best team in the SEC because of who they do play after their opening loss. I don't care about the SEC, but I thought it was interesting and have played around with it most of the season to try to get as objective a look at the results as is posslble.

I've been hesitant to share this, based on the anitcipated reaction this will get from some on the board because it hasn't painted us in the best light. But with two games to go for 16 teams (1 for Purdue and Indiana), the fluctuations aren't as wild as they were in the first part of the season. It puts some perspective on teams with similar records, but also provides perspective on how things might be fallign the way they are.

For the chart below, it's based on the current standings, but he Opponent winning percentage includes the remaing games for each team based on current records. WHile their winning percetages will change a bit, at this point over the next two weeks, it provides a snapshot and some prediction on what lies ahead. Note, this only includes the conference record.

An important note to understand the OW% - H2H column. The header stands for Opponent Winning Percentage - Head-to-head results. I calculate the openents winning percentage for all played games, then subract the results for each team from their season total to remvoe undue influences in the comparisons.

EXAMPLE
The easiest example is Purdue. The raw aggregate record of Purdue's opponents is 32 - 17. However 8 of the wins have come because they played Purdue. When comparing the winning percentage for Purdue's opponents, I subtract the wins and losses from the total. So in this case, since Purdue hasn't won a game, the record used to calculate Purdue's Opponent winning percentage is then 24 - 17.

Why do this? It's a step that looks at how those teams did against the rest of their schedule, removing the impact of their head to head resutls which inflates the opponents overall record since in this case Purdue is the worst team in terms of wins and losses in the conference. Where this really comes in handy is the teams in the middle of the pack to gauge the relative strength of schedule for teams with similar records. As you can see, not only is Purdue at the bottom, but the collective winning percentage of the teams they've played is second highest in the league at 64.3% (which won't change as they only have one remaining game, the result of which won't be included in a revised total).

The results are interesting. Having seen tOSU and Indiana play a number of games, I believe they are very good teams, the collective winning percentage of their opponents suggests that some of their success is also based in part on who they did NOT play, although in their cases, I'm not sure that would have changed their records all that much.

Because of the disparity of schedules, this doesn't prove anything, but provides some additional perspective absent head to head results with 8 of the 18 teams in the conference. The standing are ranked first by position in teh b1G then by OW% - H2H

Team	B1G Standings	OW% - H2H
Indiana	1	41.7%
Ohio State	2	39.5%
Oregon	3	51.2%
USC	3	50.9%
Michigan	3	39.5%
Iowa	6	58.1%
Illinois	6	56.8%
Nebraska	6	47.6%
Minnesota	6	46.5%
UCLA	11	52.6%
Northwestern	11	51.2%
Rutgers	13	45.6%
Wisconsin	14	64.5%
Penn State	14	58.1%
Maryland	14	46.5%
Michigan State	17	58.1%
Purdue	18	64.3%

Below is the straight rank of the opponents winning % by team highest to lowest:

 

[Ope3]

 

[MaxyJR1]

So Michigan and tOSU have had the easiest conference schedules in some sense.

 

[caliGopher]

So Michigan and tOSU have had the easiest conference schedules in some sense.

based on the way the season has played out, it appears that is accurate using the methodology I am using. Their resutls of their head to head won't figure into the final, but it will shift a little based on how next weekend falls out.

 

[Gophergrandpa]

Also would show the Badgers have played a truly tough schedule. They might be bad this year, but perhaps not as bad as everyone thinks. If the Gophrrs fail against NW, watch out.

 

[Sparlimb]

Also would show the Badgers have played a truly tough schedule. They might be bad this year, but perhaps not as bad as everyone thinks. If the Gophrrs fail against NW, watch out.

Badgers are awful offensively and pretty good defensively. Lucky for them, we make all offenses look much better then they are. Losing our last two games is definitely in the cards.

 

[Some guy]

People think this is an anomaly schedule for the gophers and the big ten

But this is the way math works

The flex protect model means there aren’t enough games to really compare teams.
Need more round robins. Either 3 groups or 2 groups or 4 groups.

 

[goonin_golden_gophers]

Interesting, although if I remember correctly this comparison introduces a performance-correlated bias [correlates with a team’s own win-loss record because the H2H adjustment rewards winning and punishes losing]. There is the Colley Matrix, which I think corrects for this in part [adjusts for opponent strength without manually subtracting head-to-head results].

May be interesting for your comparison if you have nothing better to do on a Sunday.

here is the paper: https://www.colleyrankings.com/matrate.pdf

 

[GVLCGopherFan]

So Michigan and tOSU have had the easiest conference schedules in some sense.

Ohio State never has to play Ohio State and Michigan never has to play Michigan. Wouldn’t that have impact on a stat like this? Especially since there are only eighteen teams.

 

[Some guy]

Ohio State never has to play Ohio State and Michigan never has to play Michigan. Wouldn’t that have impact on a stat like this? Especially since there are only eighteen teams.

Ohio state has played zero teams in the current top 5
Indiana has played 3 in the current top 10
Oregons second best win is Minnesota


The schedule is broken
They didn’t want divisions to make money. But I am smarter than the networks.
No divisions = no games vs teams with good records

The double bye makes this problem even worse


The problem isn’t just at the top.
It’s possible 5 teams finish with 0 or 1 conference wins
The schedule used to make that extremely unlikely to be more than 2-3 teams and IMPOSSIBLE for more than two teams to be 0-9.
You could have a year with 6 0-9 teams in the current schedule

 

[TwistingMyMelon]

I've been hesitant to share this, based on the anitcipated reaction this will get from some on the board because it hasn't painted us in the best light.

Why? Did you think that a lot of people on this board are dumb enough to believe that the Gophers this year are really good?

 

[Some guy]

Why? Did you think that a lot of people on this board are dumb enough to believe that the Gophers this year are really good?

Some people hear me say the gophers are an average college football team and get offended I called them great

But you can’t fix stupid

 

[corcorangopher]

Not sure what this ‘data’ is trying to show us. A well coached team would have beaten California and not been as embarrassed against OSU, Iowa, and Oregon. And not have left the Rutgers and MSU games up to a coin flip. Do not think that this team has over performed.

 

[Clairemont Gopher]

Y

Also would show the Badgers have played a truly tough schedule. They might be bad this year, but perhaps not as bad as everyone thinks. If the Gophrrs fail against NW, watch out.

Yup

 

[Random Guy 12]

Ohio state has played zero teams in the current top 5
Indiana has played 3 in the current top 10
Oregons second best win is Minnesota


The schedule is broken
They didn’t want divisions to make money. But I am smarter than the networks.
No divisions = no games vs teams with good records

The double bye makes this problem even worse


The problem isn’t just at the top.
It’s possible 5 teams finish with 0 or 1 conference wins
The schedule used to make that extremely unlikely to be more than 2-3 teams and IMPOSSIBLE for more than two teams to be 0-9.
You could have a year with 6 0-9 teams in the current schedule

that's why I still think divisions are better. Just add 2 more damn teams and go to 4 divisions of 5. Also makes for more meaningful games and more for teams to play for and keeps the schedule more balanced.

 

[Some guy]

that's why I still think divisions are better. Just add 2 more damn teams and go to 4 divisions of 5. Also makes for more meaningful games and more for teams to play for and keeps the schedule more balanced.

Agree

Don’t even need “divisions”
Just need scheduling groups.
These groups could change every year.

Do east west schedule one year
Do legends and leaders a year
Do old big ten new big ten one year

Etc

 

[caliGopher]

Not sure what this ‘data’ is trying to show us. A well coached team would have beaten California and not been as embarrassed against OSU, Iowa, and Oregon. And not have left the Rutgers and MSU games up to a coin flip. Do not think that this team has over performed.

When I started this exercise, the goals was to get some sense of how to consider the results of the relative strength of the teams to each other when any one team plays 9 conference games, not playing the other 8. So it's a strength of schdule of a sort.

As the year went on, I tweaked it when looking at each team's record because of Purdue, Michigan State, WIsconsin, and Penn State's poor records as the season unfolded to look only at the record of their opponents without their results included. I did that to understand, how did the opponents do when not playing the team the score was cacluated. Leaving tOSU's 7 wins in the calculation would deflate the relative strength of it's opponents while Purdue's 8 losses would inflate it.

I don't look at this and say, team A has under preformed or team B has over preformed, but Wisconsin's struggles seem to be far easier to understand, than say, Rutgers.and of the teams knotted at 4-3, Iowa and Illinois appear to be stronger 4-3 teams than the rest.

 

[caliGopher]

Interesting, although if I remember correctly this comparison introduces a performance-correlated bias [correlates with a team’s own win-loss record because the H2H adjustment rewards winning and punishes losing]. There is the Colley Matrix, which I think corrects for this in part [adjusts for opponent strength without manually subtracting head-to-head results].

May be interesting for your comparison if you have nothing better to do on a Sunday.

here is the paper: https://www.colleyrankings.com/matrate.pdf

Thanks for sharing this. This was interesting to read through and I appreciate your bringing it up. Admittedly, my attempts were not nearly as in depth as Dr. Colley's.

I toyed with redoing this using his formula, but did not for two reasons.

1) My goal was not to rank the teams in total as his process does, it was just to provide myself some perspective in a league where we don't play roughly 50% of the teams in a given year.

2) When I read more into his matrix and results, I gave up the idea completely given that his model of predicting the top team "without bias"only strayed from the eventual champion 4 times in 27 years. (the most of any of the polls in the BCS/Playoff era) and when it did....

1. in 2011, Colley reanked Oklahoma State as the best team, but they did not play in the title game (this is the only one that an argument could be made in defense of the difference)
2. in 2012 Colley ranked Notre Dame ahead of Alabama in the final poll even though Alabama beat Notre Dame in the BCS Championship game 42-14
3. in 2016, Colley ranked Alabama ahead of Clemson in the final poll even though Clemson beat Alabama in the Championship game
4. in 2017 Colley ranked UCF as #1.

 

[Scottsdale Gopher]

The problem isn’t just at the top.
It’s possible 5 teams finish with 0 or 1 conference wins
The schedule used to make that extremely unlikely to be more than 2-3 teams and IMPOSSIBLE for more than two teams to be 0-9.
You could have a year with 6 0-9 teams in the current schedule

It's been mentioned elsewhere that the size of the conferences now creates a vast middle that frequently will include teams that do not consider themselves part of the middle. But the other thing the vast middle creates is a lot of games between teams of similar levels of talent. I think the Gophers are a great example of that this year. I would not be surprised if they finish 0-2, but finishing 2-0 for a 6-3 conference record would be a reflection of them going undefeated against similarly talented teams this season.

Cue Paul Harvey: Best of the lousiest, lousiest of the best, etc.

 

[Random Guy 12]

It's been mentioned elsewhere that the size of the conferences now creates a vast middle that frequently will include teams that do not consider themselves part of the middle. But the other thing the vast middle creates is a lot of games between teams of similar levels of talent. I think the Gophers are a great example of that this year. I would not be surprised if they finish 0-2, but finishing 2-0 for a 6-3 conference record would be a reflection of them going undefeated against similarly talented teams this season.

Cue Paul Harvey: Best of the lousiest, lousiest of the best, etc.

Where does Iowa fit into that theory? I would say purgatory. Can't quite get to the level of the elite but better than the middle? Realistic shot that we end up with a better record than they do however.

 

[The Deliberate Stranger]

So in short we are 2-0 against the bottom of the league, 1-0 against whatever we want a call slightly better than the bottom of the league, 1-1 against teams with the same record (Are we Iowa's best win?) 0-2 against the top of the league. We are in a five way tie for sixth place where our win in that group was at home and the loss was on the road. The teams at the top of the league have the "easiest" schedule and the teams at the bottom of the league have the "hardest schedule" Real earth shattering stuff!

 

[Scottsdale Gopher]

Where does Iowa fit into that theory? I would say purgatory. Can't quite get to the level of the elite but better than the middle? Realistic shot that we end up with a better record than they do however.

I think Iowa is roughly what most of this board wants/believes PJ provides: a half-step to a full-step above the middle with an opportunity every five years or so to be amongst the elite. But in a big conference with unbalanced schedules, you might find your "peers" to be lesser teams.

 

[Some guy]

It's been mentioned elsewhere that the size of the conferences now creates a vast middle that frequently will include teams that do not consider themselves part of the middle.

Disagree completely.
The conference is very stratified in thirds.

The top is bigger and the bottom is bigger. The middle is smaller.
The top doesn’t play each other keeping more 0-2 loss teams than usual.
The bottom doesn’t play each other keeping more 0-2 win teams than usual.

But the other thing the vast middle creates is a lot of games between teams of similar levels of talent. I think the Gophers are a great example of that this year. I would not be surprised if they finish 0-2, but finishing 2-0 for a 6-3 conference record would be a reflection of them going undefeated against similarly talented teams this season.

Agree with that. But it’s not because of the “vast middle” it’s because the gophers didn’t play the middle. They played 2 teams in top 3rd
2 teams in the middle third.
5 teams in bottom third. (If they beat northwestern)

Cue Paul Harvey: Best of the lousiest, lousiest of the best, etc.

I half agree with this quote.
It is quite possible they are the best of the lousiest. They’re far away from being in a tier with the best.

 

[Scottsdale Gopher]

Agree with that. But it’s not because of the “vast middle” it’s because the gophers didn’t play the middle. They played 2 teams in top 3rd
2 teams in the middle third.
5 teams in bottom third. (If they beat northwestern)

My concept is that the tiers are talent based, not record based. The distribution of talent in B1G is very condensed (as seen in annual recruiting rankings), such that there are not true thirds as you're suggesting. Here's the team talent rankings for the B1G from cfbdepth.com with their national rankings in parentheses.

1. Ohio State (1)
2. Indiana (2)
3. Oregon (3)
4. USC (15)
5. Iowa (19)
6. Michigan (20)
7. Washington (28)
8. Illinois (31)
9. Rutgers (33)
10. Nebraska (39)
11. Minnesota (46)
12. Northwestern (54)
13. Penn State (58)
14. Michigan State (61)
15. Maryland (64)
16. Purdue (96)
17. UCLA (102)
18. Wisconsin (128)

I don't necessarily agree with the specific rankings, but the shape of it is consistent with what I'm saying. There are three schools more than 1 standard deviation (~36) higher than the average ranking (44) and three schools more than 1 standard deviation below the average ranking. A dozen schools (USC - Maryland) reside in what I called the "vast middle", but is the center of the bell curve. We could choose to split the middle around the mean (four total groups), but it wouldn't really change our conversation about the Gophers as they are effectively the mean and they have played 2 teams in the top tier and will play 2 teams in the bottom tier. Every other opponent on the conference schedule was within one standard deviation of their talent level.

If the Gophers go 2-0 to end the season, they'd be 0-2 against the most talented B1G teams, 4-1 against the middle, and 2-0 against the least talented B1G teams.

And yes, the Paul Harvey comment was just for fun.