Statistical Strength of Schedule

Note: This is the first post in the SSOS series. Check the main page by clicking the banner above to see the most recent SSOS numbers.

I have created a simple statistical method of measuring strength of schedule. I don’t like using just wins and losses for a couple of reasons. For one, a win over Louisiana-Monroe counts the same as a win over Louisiana State. Beyond that, just look at the Florida – LSU game. Florida led for 58:51 of that game, but in a wins and losses only system, Florida gets no credit for that.

So, I put everyone’s schedule into four different spreadsheets and filled in each team’s opponents’ rank in total offense, scoring offense, total defense, and scoring defense. I then calculated the average rank of each that each team has faced, then averaged those four averages to come out with a final strength of schedule number. Now that everyone has played 5 or 6 games, the sample size is large enough that most anomalies should be evened out by now.

As a side note, the rank for any I-AA team in any category is 120. I am using the official NCAA stats, and they include games against I-AA teams. I do not have the time to go through everyone’s games to subtract out all I-AA games from the stats and recalculate the categories. I also didn’t want to just lump all of Division I together like Jeff Sagarin does because I-AA teams play other I-AA teams almost exclusively, and the better I-AA teams would end up a lot higher than they should be. Instead, I just gave a blanket penalty to I-A teams for playing I-AA teams because they should know better.

For these purposes, Western Kentucky is still a I-AA team despite being a I-A provisional team. Here’s the top 25:

  1. Notre Dame
  2. Stanford
  3. Washington
  4. Akron
  5. Auburn
  6. Florida International
  7. East Carolina
  8. Tennessee
  9. Nebraska
  10. Ole Miss
  11. BYU
  12. Minnesota
  13. Marshall
  14. Washington State
  15. Colorado
  16. Texas
  17. UL-Monroe
  18. NC State
  19. Syracuse
  20. Duke
  21. Utah State
  22. Colorado State
  23. Oregon State
  24. Buffalo
  25. Alabama

The whole, color-coded table in PDF format is here: 10-06-07.pdf

You would expect to see bad teams in the top quintile since they do tours of BCS leagues to get paychecks, and there are some of those with Marshall, Louisiana-Monroe, and the like. However, there are also four Pac 10 teams, three SEC teams, three Big 12 teams, two ACC teams, and one each from the Big East and Big Ten.

You would also expect to see BCS league teams in the bottom quintile since they are notorious for fattening up on cupcakes in the early season. To that end, we find five Big 12 teams, three Big Ten teams, three Big East teams, and one each from the Pac 10 and SEC. Those teams from the Pac 10 and SEC? USC and Vanderbilt. For all the complaining about SEC teams and their non conference scheduling, from a statistical standpoint, only Vandy has a deplorable schedule. It also shows that so far, Les Miles was right about USC having an easy road, though all of the Trojans’ toughest games are still ahead.

I then went through and arranged everything by wins and losses, and some surprising and not-so-surprising things surfaced. Of the eleven undefeated teams left, five have schedules in the bottom quintile, three in the fourth quintile, two in the third quintile, and only one – LSU – from the second quintile. Not only have everyone in the top quintile lost a game, but all have lost at least two games. I think that helps validate this method some. It also shows just how good LSU is for being undefeated against the 41st toughest schedule when the next highest unbeaten is Cal at 56, more than half a quintile below.

The worst schedule by far is Hawaii’s, with two division I-AA teams plus Louisiana Tech, UNLV, Idaho, and Utah State. There are about 9 SOS points separating Hawaii from the second-worst, UConn, and for reference, there are about 9 SOS points separating #30 Miami of Ohio (60.25) from #76 Rutgers (69.55). At least Hawaii was smart by putting Boise State and Washington at the end, because if the Warriors can beat those two teams, the voters will only remember those two games and not the 10 other wretched teams on the schedule and grant them spot in a BCS bowl, where they will likely get unceremoniously blown out.

Here are some more fun stats that come out of these numbers.

Average SOS by Record

This actually plays out how you would expect. The better the record, the worse the strength of schedule.

There is one buck to the trend, and that’s 2-loss teams ending up lower than you would think. Well, the category of two-loss teams includes a lot of BCS conference schools like Florida, Auburn, Kansas State, and Texas, who have played tougher schedules than some of those around them since they have had early-season in-conference tests. In other words, had they played easier schedules, they’d probably not have lost two games.

Total Average SOS for all 119 Teams: 66.48

6-0: 77.42

5-0: 77.29

5-1: 71.58

4-1: 67.54

3-1: Only Boise State is listed at 3-1 since the Broncos already had their bye week and their Sunday night game was not counted in this week’s stats. Their SOS is 71.85, as listed above.

4-2: 64.58

3-2: 64.93

3-3: 67.27

2-3: 66.53

2-4: 62.42

1-4: 63.75

1-5: 63.21

0-5: 56.80

0-6: 52.03

SOS by Conference

Best Schedule: Stanford, 2nd overall, score of 42.45

Worst Schedule: USC, 108th, 79.60

Average SOS Rank: 42.4

Average SOS Score: 60.31

Best Schedule: Auburn, 5th overall, score of 46.92

Worst Schedule: Vanderbilt, 98, 61.91

Average SOS Rank: 44.25

Average SOS Score: 77.00

Best Schedule: NC State, 18th overall, score of 55.21

Worst Schedule: (tie) Boston College and Clemson, 95th, 76.04

Average SOS Rank: 61.00

Average SOS Score: 66.35

Best Schedule: Nebraska, 9th overall, score of 51.33

Worst Schedule: Kansas, 116th, 86.50

Average SOS Rank: 63.67

Average SOS Score: 68.38

Best Schedule: Minnesota, 12th overall, score of 53.88

Worst Schedule: Indiana, 107, 78.71

Average SOS Rank: 71.45

Average SOS Score: 69.28

Best Schedule: Syracuse, 19th overall, score of 66.63

Worst Schedule: UConn, 118th, 91.45

Average SOS Rank: 81.25

Average SOS Score: 73.76

As you can see, it’s roughly a dead heat between the Pac 10 and SEC for scheduling, so if any of you SEC fans out there catch any heat from Pac 10 fans or especially Big Ten fans about the SEC scheduling weak teams, you can now set them straight.

Advertisements

5 Responses to Statistical Strength of Schedule

  1. bcsbusters says:

    This is Outstanding Stuff! I was wondering how to put all of this together for I feel the same way as you. And how do you think I feel about OU being ranked 5th in country when Colorado is not only the better team on the field, but has played a tougher schedule. People bark about USC remaining in the Top-10, but no one says a word about Oklahoma?

    Would you mind if I posted this on my blog and also on the Bleacher Report, with a link and reference to you of course!

  2. bcsbusters says:

    Of course if you took Division I-AA out of the equation (2 PAC-10 teams and I believe 7 SEC teams), I don’t think the numbers come out quite the way you have predicted. Nonetheless, although people have been raving about the PAC-10 being up this year – I actually think it is running about average, I think the SEC is slightly ahead, but only because you have 12 teams instead of 10 and you have more stronger teams in the 4-8 slots than the PAC on an annual basis. Of course I’ve always laughed about the absurdity of the PAC as a weak conference considering they have won more national championships than all of the other conferences combined across the board, and have also had a team worthy of BCS championship consideration throughout the BCS era – USC of course, but then UW, OSU, Oregon and even UCLA was in the hunt a few years ago.

  3. year2 says:

    You may republish whatever you want as long as you link it like you said.

    I wish there were stats of just I-A versus I-A teams, but I can’t find any as easily accessible and manipulable as the stats from the NCAA which include I-A versus I-AA games.

  4. bcsbusters says:

    Got to thinking about your stats – which I really like by the way, but one reason why the elite teams are ranked so low is they don’t have to factor in playing themselves. I think we should include every teams statistical components into their own SOS (their opponents as well) and recompute the numbers. I might make for a more accurate assessment. Let me know your thoughts and if you could crunch those numbers for me I would appreciate it.

  5. year2 says:

    Well, strength of schedule is only about how difficult the opponents you have faced are. Oklahoma is going to do more for Colorado’s SOS than Colorado will do for Oklahoma’s; USC does more for Stanford than vice versa and so on. That’s just the nature of the beast.

    What’s hurting the top teams is that a lot of them still have the toughest part of their schedule ahead of them. For instance, Ohio State is low because it hasn’t played Wisconsin, Illinois, Michigan State, or Michigan yet. The front-loaded schedulers are doing well now, and the back-loaded schedulers are suffering. It will all even out in the end if a team is indeed playing a legitimately tough slate of games.

    If you’re interested in the interplay between own performance and SOS, check out my post called SSOS Power Poll:
    https://year2.wordpress.com/2007/10/11/ssos-power-poll/

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: