How Important Is Strength of Schedule Based on Last Year?

First, there are different ways to measure schedule strength. Most outlets will simply use the previous year’s records for the upcoming opponents. I’ll do that here to compare, but generally, the better info you get, like underlying point differentials or per play efficiency numbers, the better you will be. Adjusting for schedule strength (in essence, opponent’s opponent’s records) matters too. But for this post, I’ll simply compare the previous year’s records to what happened.

Over the last four years, the year to year win total correlation in the NFL has been +0.35. That has bounced around from a low of +0.21 for 2007-2008 to +0.57 for 2008-2009. What does that mean? It means that there is some relationship between knowing last year’s win total and the next year’s, but it’s far from strong. Some good teams stay good, some bad teams stay bad. It’s not simply redrawing names from a hat.

I looked at schedules for the last four years, what the projected strength of schedule would be, and then the actual results. You would think that with wins not being perfectly correlated already, that the randomness of who improves or declines would destroy any correlation between projected and actual schedule strength even more. At least I did.

Not so. The correlation coefficient between projected strength of schedule and actual strength of schedule is +0.32, not that much different from the year to year correlation in wins. Interestingly (at least to a geek like me) is that the correlation between projected and actual strength of schedule was highest in the two years where individual team year to year win correlation was lowest.

I wonder if this has to do with the current scheduling, where teams play blocks of other teams. We may not know which teams will rise and fall, but within a group, its bound to happen. Thus, when one team rises within a conference, there is probably another one you are also playing who declined.

Anyway, here is a quick chart showing the average strength of schedule for teams from the last four years, subdivided into five groups based on projected schedule strength.

Not much difference, except at the extremes. Even then, we are talking about a 9 win difference in actual results from the projected hard schedules and projected easy ones (roughly the difference of one more game against a playoff team versus a bottom 5 team). None of the 10 easiest projected schedules were easier than thought. None of the 10 hardest ones turned out to be as hard as expected.

Does preseason strength of schedule matter? Yes and No. It matters in that I would rather play a projected easy schedule, yes. I wouldn’t devote too much time to the minutae of each team on the schedule, though.

Just for fun, I’ll close with a couple of notes on the teams who have benefitted the most and least from the year to year changes in the schedule. Over the last four years, every team in the AFC has played every team in the NFC, but catching a division in one year versus another could constitute scheduling luck. Add in those conference games where catching the right team that declines rapidly also affects your schedule strength, and here are the teams that have benefitted the most.

1. Indianapolis (projected schedule strength 0.535, actual 0.496)

2. New Orleans (projected schedule strength 0.496, actual 0.458)

3. Jacksonville (projected schedule strength 0.532, actual 0.497)

4. Tennessee (projected schedule strength 0.521, actual 0.492)

5. Houston (projected schedule strength 0.529, actual 0.500)

This suggests the AFC South has tended to catch other divisions in the right year for the last four seasons, making their schedules easier than they appeared entering the year. On the other end of the spectrum, poor Rams.

1. St. Louis (projected schedule strength 0.470, actual 0.523)

2. Cleveland (projected schedule strength 0.511, actual 0.546)

3. Detroit (projected schedule strength 0.509, actual 0.540)

4. Oakland (projected schedule strength 0.478, actual 0.505)

5. Seattle (projected schedule strength 0.469, actual 0.493)

We don’t see one division dominating on this end, which suggests these teams caught the wrong teams in the conference matchups that are based on finish. The Rams have not only been bad, they haven’t caught the schedule breaks either.

[photo via US Presswire]