Is Strength of Schedule worth worrying about?

Strength of schedule is one of the many factors that fantasy football players often include in their player rankings to try and get an edge over their opponents. One of the problems with strength of schedule however is that there is so much parity in the NFL that a bad run defense can become one of the best ones in the following year. This article will consider statistics from the past four seasons and will try and determine whether or not strength of schedule is one of the factors worth worrying about in fantasy football.

The first step for this analysis was to determine how much defensive statistics vary from year to year. In order to do this I used the “points allowed by position” statistics which tell us how many fantasy points each team allowed to quarterbacks, running backs and wide receivers in a given season. The goal was to find the degree of correlation between the points each team allowed in two consecutive seasons. Using statistics from 2003 through 2007 here are the formulas that predict the most accurately each team’s points allowed by position for the next season:

	Formula
QB	Next Year Pts Allowed = 214 + 0.223 x Previous Year Pts Allowed
RB	Next Year Pts Allowed = 303 + 0.196 x Previous Year Pts Allowed
WR	Next Year Pts Allowed = 400 + 0.230 x Previous Year Pts Allowed

These formulas show us that it is very difficult to predict how defenses will perform in the upcoming season as only 20% to 25% of the previous year’s statistics help us predict next year’s statistics. For example, the Cleveland Browns were the worst passing defense last season by allowing 333 fantasy points to opposing quarterbacks but according to the above formula you can expect them to allow only 288 in 2008. For those who doubt these formulas, of the ten defenses that allowed the most fantasy points to quarterbacks in 2006 only two remained in the bottom ten in 2007 and three were in the top ten.

The next step was to use these numbers and determine each team’s strength of schedule by position for the 2008 season. The following charts show the three easiest and the three hardest schedules for each position as well as the multiplier which shows how much easier or harder their schedule is compared to the league average:

Easiest Schedules
QB		RB		WR
WAS	1.008	MIN	1.017	WAS	1.016
PHI	1.007	CAR	1.015	SF	1.015
TB	1.007	NO	1.015	NYG	1.009

Hardest Schedules
QB		RB		WR
SD	0.989	CIN	0.980	SD	0.985
NE	0.991	NYG	0.983	NO	0.988
DEN	0.993	PIT	0.983	DEN	0.989

These numbers show us that the difference is less than 4% between the teams with the easiest schedules and the teams with the hardest schedules. For example the Minnesota Vikings have the easiest schedule for running backs but we can expect the defenses they will face to allow only 1.7% more fantasy points to running backs than the league average. This means that if you project Adrian Peterson to have 300 fantasy points, his easier schedule would mean you can expect him to have 305 fantasy points. There are so many factors in fantasy football and fantasy players try and include as many as possible in their rankings but I believe these statistics show that projected strength of schedule is one that is almost negligible.

There is one factor that is rarely looked at that I believe is much more important than projected strength of schedule and that is the previous season’s strength of schedule. If you think about it, it does not make much sense that it is more common to try and predict how good a player’s opponents will be next season than looking at how good his opponents were last season. The following charts are very similar to the ones shown previously but in this case they will show the three teams who had the easiest and hardest schedules in 2007 based on 2007 statistics:

Easiest Schedules
QB		RB		WR
PIT	1.055	MIN	1.086	SEA	1.081
TB	1.045	GB	1.072	SF	1.078
ARI	1.040	KC	1.060	PIT	1.058

Hardest Schedules
QB		RB		WR
DEN	0.940	BUF	0.936	DEN	0.947
JAX	0.952	NYJ	0.938	SD	0.950
NO	0.959	CIN	0.946	TEN	0.951

As you can see these numbers vary much more than the ones shown previously and it is quite normal because instead of applying the formulas that only take into account 20% to 25% of the 2007 statistics we just used the 2007 statistics. The difference is quite remarkable between the teams with the easiest schedules and the teams with the hardest schedules. Ben Roethlisberger for example had 306 fantasy points but he faced defenses that allowed on average 290.1 fantasy pts while the league average is 274.9. If you divide 306 by 1.055 you obtain 290 fantasy points, Ben Roethlisberger’s estimated number of fantasy pts in 2007 had he faced average defenses. If you do the same thing for Jay Cutler who had 267 points with the hardest schedule for quarterbacks, his adjusted number of fantasy pts for 2007 is 284. This means that Roethlisberger who seemingly was much superior to Cutler last season in fact only had six more points than him if we adjust for strength of schedule.

Now these multipliers are not perfect because some would argue that for example Minnesota’s opponents allowed more points to running backs than the average defense because they all had to face the Vikings at least once. It is true that a run offense that was as dominant as Minnesota may have a strength of schedule multiplier that is a little too high but I do not think it has a significant impact. These numbers are not directly used in Ultimate FF Strategy’s 2008 Projections but they might help explain a few things like the very solid seasons of Kurt Warner, Adrian Peterson and Ryan Grant or the disappointing seasons of Thomas Jones and Rudi Johnson. The reason these numbers are not used directly in our projections is that the historical statistics we look at were not adjusted for strength of schedules either. For example, when we say that players who had a yards per carry average between 5.2 and 6.0, like Peterson, had a decrease of 18% in YPC in the next year, part of the reason for that 18% decrease for many players could have been that they had an easy strength of schedule. It would therefore be incorrect to decrease Peterson’s projected yards per carry average by 15% to 20% because similar players had that decrease and then by another 8% because he had an easy schedule last season. For those interested we have posted the complete 2007 strength of schedules by position.

Defenses change so much from year to year that it is not very useful to look at a player’s strength of schedule based on the previous year’s defensive statistics. Certain people do try to adjust defensive statistics based on changes made in the off-season but it is such a small factor that you are much better spending your time on predicting the offensive players’ statistics. However, looking at the previous year’s strength of schedules to adjust player statistics will allow you to compare players much more accurately and could help you explain a few things that will give you an edge over other fantasy football players in your draft.