Strength of Schedule: What is it Good For?

(Note: Check out this article at our new location with a brand-new look and feel… Blog-a-Bills)

Absolutely Nothing?

Every year during this time, before preseason games have begun but after training camp has started, optimism abounds in many NFL cities (with the possible exception of cities that may or may not rhyme with Muffalo). We, as fans, breathlessly await word of how draft picks perform in their first taste of any kind of NFL action, eagerly anticipate news of the latest step in a recovery from injury, and parse every training-camp related Tweet to hear about any training camp surprises (like Stevie Johnson).

Finally, many of us parse the year’s upcoming schedule, guessing at wins and losses during each week, in an attempt to predict how many wins are possible. It is during this process we most often pay attention to the “strength” of our team’s schedule. As Bills fans, many of us undoubtedly have heard or read that the Bills have the 2nd hardest schedule in the league this year, which may affect the level of optimism we have about this years’ team.

All this begs the question: How much can we actually tell about a team’s chances in the upcoming year from strength of schedule alone?

Before we delve into this question, a few quick notes on the information used in this study:

1. Strength of schedule is defined here as it is when the NFL first releases schedules: a compilation of the previous year’s win/loss records for all the team’s opponents in the upcoming year.
2. The 2010, 2009 and 2008 seasons are included in this analysis.
3. In the tables below, Wins refers to how many actual wins that team picked up that year, while point differential refers to the cumulative point differential for a team that year.

2010 Data

Rank	Team	Opp W	Opp L	Opp Win %	Wins	Point Diff
1	Texans	140	116	0.547	6	-37
1	Titans	140	116	0.547	6	17
3	Cowboys	139	117	0.543	6	-42
4	Bengals	138	118	0.539	4	-73
5	Jaguars	137	119	0.535	8	-66
6	Patriots	136	120	0.531	14	205
7	Giants	135	121	0.527	10	47
8	Redskins	134	122	0.523	6	-75
9	Eagles	133	123	0.52	10	62
10	Browns	132	124	0.516	5	-61
10	Colts	132	124	0.516	10	47
12	Ravens	130	126	0.508	12	87
12	Lions	130	126	0.508	6	-7
14	Bears	129	127	0.504	11	48
14	Vikings	129	127	0.504	6	-67
16	Bills	128	128	0.5	4	-142
16	Dolphins	128	128	0.5	7	-60
16	Jets	128	128	0.5	11	63
16	Raiders	128	128	0.5	8	39
20	Falcons	127	129	0.496	13	126
21	Steelers	126	130	0.492	12	143
22	Packers	125	131	0.488	10	148
22	Chiefs	125	131	0.488	10	40
24	Broncos	124	132	0.484	4	-127
25	Buccaneers	123	133	0.48	10	23
26	Panthers	122	134	0.477	2	-212
27	Saints	120	136	0.469	11	77
28	49ers	117	139	0.457	6	-41
29	Chargers	116	140	0.453	9	119
29	Seahawks	116	140	0.453	7	-97
31	Rams	115	141	0.449	7	-39
32	Cardinals	114	142	0.445	5	-145

2009 Data

Rank	Team	Opp W	Opp L	Opp Win %	Wins	Point Diff
1	Dolphins	152	104	0.594	7	-30
2	Panthers	151	104	0.592	8	7
3	Patriots	151	105	0.59	10	142
4	Falcons	150	105	0.588	9	38
5	Buccaneers	148	107	0.58	3	-156
6	Bills	146	110	0.57	6	-68
7	Jets	145	110	0.568	9	112
8	Saints	142	113	0.557	13	169
9	Eagles	137	119	0.535	11	92
10	Giants	134	120	0.527	8	-25
11	Cowboys	131	123	0.516	11	111
12	Jaguars	132	124	0.516	7	-90
13	Colts	131	125	0.512	14	109
14	Titans	130	126	0.508	8	-48
15	Texans	129	126	0.506	9	55
16	Redskins	125	129	0.492	4	-70
17	Chiefs	123	131	0.484	4	-130
17	Chargers	123	131	0.484	13	134
19	Broncos	122	132	0.48	8	2
19	Raiders	122	132	0.48	5	-182
21	Lions	119	136	0.467	2	-232
22	Bengals	119	137	0.465	10	14
22	Rams	119	137	0.465	1	-261
24	Seahawks	117	139	0.457	5	-110
25	Browns	114	140	0.449	5	-130
26	49ers	113	142	0.443	8	49
27	Cardinals	113	143	0.441	10	50
28	Ravens	111	143	0.438	9	130
29	Steelers	110	144	0.434	9	44
30	Packers	109	146	0.428	11	164
31	Vikings	107	148	0.42	12	158
32	Bears	105	149	0.414	7	-48

2008 Data

Rank	Team	Opp W	Opp L	Opp Win %	Wins	Point Diff
1	Steelers	153	103	0.598	12	124
2	Colts	152	104	0.594	12	79
3	Jaguars	143	113	0.559	5	-65
4	Vikings	141	115	0.551	10	46
4	Ravens	141	115	0.551	11	141
6	Bengals	140	116	0.547	4	-160
6	Texans	140	116	0.547	8	-28
6	Browns	140	116	0.547	4	-118
9	Lions	139	117	0.543	0	-249
9	Titans	139	117	0.543	13	234
11	Bears	136	120	0.531	9	25
11	Packers	136	120	0.531	6	39
13	Redskins	134	122	0.523	8	-31
13	Cowboys	134	122	0.523	9	-3
15	Eagles	133	123	0.52	9	127
15	Giants	133	123	0.52	12	133
16	Rams	125	131	0.488	2	-233
17	49ers	124	132	0.484	7	-42
18	Seahawks	122	134	0.477	4	-98
19	Buccaneers	120	136	0.469	9	38
20	Cardinals	119	137	0.465	9	1
20	Dolphins	119	137	0.465	11	28
20	Panthers	119	137	0.465	12	85
24	Falcons	118	138	0.461	11	66
25	Jets	117	139	0.457	9	49
26	Chiefs	116	140	0.453	2	-149
27	Bills	115	141	0.449	7	-6
27	Saints	115	141	0.449	8	70
29	Broncos	114	142	0.445	8	-78
30	Raiders	112	144	0.438	5	-125
31	Chargers	108	148	0.422	8	92
32	Patriots	99	157	0.387	11	101

Enough With the Tables – Is Strength of Schedule a Reliable Predictor of a Team’s Record?

In a word:

No.

In more words:

Intuitively, one would assume that the more losses a team’s opponents had the year before (i.e. an “easier” schedule ) would be positively correlated with the number of wins the team in question had the next season (i.e. the more losses the Bills opponents had last year would have some positive relationship with number of wins the Bills would have this year. Well, maybe not the Bills, but some NFL team).

However, after running a correlation analysis in Excel between a team’s strength of schedule and their wins for that year, we found that there is virtually no connection at all between the two. For the three years worth of data used in this study, the correlation was a mere -0.0457. (Correlations can run from -1 to 1, with a value of 0 meaning there is absolutely no correlation between two variables.)

-0.0457 is so close to 0 that there is no case to be made for strength of schedule having ANY relationship to how many wins a team produces in a given year. As a point of comparison, there are stronger correlations (but still not substantive) between a team’s point differential from the previous year and their record in the upcoming year (0.386) and between the number of wins a team had the previous year to the number of wins in the upcoming year (0.335).

To put it bluntly, we would come up with more accurate projections of the Bills 2011 record using their 2010 point differential or their 2010 winning percentage than we would by using their strength of schedule. (Of course, neither of those stats bodes well for the Bills’ chances in 2011 either).

Playoff Teams vs Non-Playoff Teams

At this point you may be saying to yourself, “ok, that’s fine, but aren’t some teams *cough* the Bills *cough* so consistently bad that they may skew these correlations? In other words, don’t the bad teams always performing poorly against everyone weigh down this comparison?” And that’s certainly a fair enough question to ask, if not also an unneeded (but accurate) cheap shot at the Bills.

In order to account for that possibility, we broke out the data and separated non-playoff teams from playoff teams. We then ran the same correlation analyses used earlier for each group of teams to see if either group had a stronger correlation, in addition to seeing if there was actually any difference in the total strength of schedule between playoff teams and non-playoff teams.

The results did not differ significantly from our earlier analysis. The cumulative strength of schedule for non-playoff teams over the past 3 years (expressed as win percentage) was 0.4995. Playoff teams over the past 3 years faced opponents whose win percentage cumulatively was 0.502. Further analysis in Excel shows that there is no meaningful difference between those two numbers; i.e. they are, statistically speaking, equivalent. In other words, there was no difference in the strength of schedules between non-playoff teams and playoff teams over the past 3 years.

The correlation analyses turned up some more interesting information. Statistically questionable, but interesting. For the non-playoff teams over the past 3 years, the correlation between strength of schedule and wins was as low as it was when looking at teams as a whole. However, playoff teams did have a stronger (but still weak) correlation of -0.292. Of course, the interesting part about that is that it implies that the “stronger” the schedule playoff teams faced between 2008-2010, the more wins they had. Given the sample sizes when looking at just playoff teams, it may be best to anoint this as an aberration, and more importantly, we need to remember that (I can’t believe I’m actually writing this term non-ironically) correlation does not equal causation. But this is still interesting nevertheless.

Teams Above .500 vs Teams Below .500

Of course, there is no guarantee that teams with winning records make the playoffs. Between arcane tiebreakers, the continued insistence on awarding playoff spots to division winners (which did give us all one more chance to see Marshawn in Beast Mode running over defenders and not pedestrians), and other vagaries in the system, there are plenty of examples of mediocre teams making the playoffs when perhaps others seemed more deserving based on merit. In the past 3 years alone there have been two 10 win teams and one 11 win team who haven’t made the playoffs.

So, we decided to take another look at the data, this time separating teams who finished .500 or below from teams who finished better than .500, and then running through the same exercises we did when looking at playoff teams vs. non-playoff teams.  Once again, after crunching those numbers, we found that there was no appreciable difference between the two groups in how difficult their schedules were, nor did we find any meaningful correlation between strength of schedule and wins in either group.

The overall winning percentage of the opponents of the teams who finished .500 or below turned out to be 0.4997, whereas the overall winning percentage of opponents of the teams who finished above .500 was 0.5018. Again, this difference is so small as to be statistically insignificant. And, as we found before, there were no significant correlations within either group between strength of schedule and wins.

One More Way to Divide the Teams

To this point, we’ve looked at team success as being defined by either playoff berths or finishing above .500. There is another stat useful in determining how a team performed during a year, though, and that is their overall point differential. For example, from 2008-2010, the Bills have been outscored by a total of 216 points: 142 in 2010, 68 in 2009, and 6 in 2008. (Sidenote for Bills fans: It could be worse. The Panthers were outscored by 212 points in 2010 alone.) This stat of course is not perfect in and of itself as it ignores garbage time scores from either team, but statistically is usually a decent indicator of a team’s competence (or lack thereof).

At this point, it is not surprising that these results did not prove to be any different than the results from those using playoff berths or total wins as a measure of success. There was no statistical difference to be found in the strengths of schedules between those who had positive point differentials and those who had negative point differentials, and no correlation was found between strength of schedule and actual results for either group.

What the Hell Does Any of this Mean, Really?

The only real conclusion that we can take away from this exercise is that the way the NFL currently defines strength of schedule (prior to the season starting and using opponent’s previous year’s records) has absolutely no bearing whatsoever on how a team can be expected to perform.

This is not totally surprising as the current definition of strength of schedule is dependent on the opponents’ combined W-L from the previous year, and as we all know (and as the NFL likes to remind us), the NFL is a league of parity. One year’s last-place division finisher could be (and has been) next year’s Super Bowl winner. Records from year to year (unless you’re the Patriots or Steelers) tend to be so volatile that there really is no rhyme or reason to using strength of schedule for any reason at all.

In the end, to Bills fans, this gives us one less reason to fret (which puts as at about 999,999 things to worry about), as the fact that the Bills have the 2nd “hardest” schedule this year means zilch when it comes to how they do this season. Unfortunately, Chris Kelsay being listed as a starter does, but that’s another topic for another day….