Fantasy: Using PFF Premium Stats to Win Your League, Part 2

This article is the second installment in an ongoing series examining how the myriad Premium Stats provided by PFF can be used to your advantage when assembling your fantasy team. In Part 1, we learned how team-level offensive statistics can be used to predict an offense’s collective fantasy output the following season. Part 2 applies similar team-level analyses to the defensive side of your fantasy roster.

I suppose I should begin with the following confession: I’m no IDP expert. In fact, the last fantasy league of mine to require any defensive players at all was also my first league, and it required only 3. Oh, and did I mention that this was eleven years ago?^[1] My lack of enthusiasm for IDP leagues probably stemmed from the fact that, at the time my interest in fantasy football was piqued, there just weren’t a lot of IDP resources available to fantasy players. Most major fantasy publications had, at best, token analyses of defensive players. If you were lucky, you could maybe even find a cheat sheet or two. As a player who likes to have a stack of expert opinion on hand before making draft-day decisions, I just couldn’t get too excited about what amounted to a dartboard approach to building a defensive squad.

Well, a lot has changed over the past decade, and it’s probably high time I admit to myself that these IDP prejudices are outdated. Most fantasy football outlets now have multiple analysts dedicated to IDP coverage, and Pro Football Focus is no exception. In fact, the IDP staffers here at PFF, led by Jeff Ratcliffe and Ross Miles, provide a quality and depth of defensive coverage that any site would be hard-pressed to match.

This is really just a long-winded way of saying: (1) I’m not here to offer an opinion on how early you should consider drafting Von Miller, or how I think Stephon Gilmore will fare in the Bills’ secondary, and (2) there are some really qualified folks out there who can answer those exact questions. What I can offer you, however, is a macro-level view of the relationship between PFF’s complex statistics and defensive fantasy performance. The main advantage of being a PFF subscriber is the ability to access PFF’s Premium Stats. But what good is any stat about a player or a team, Premium or otherwise, if you can’t use that information to gain an edge over your league mates?

That’s where I come in. This week’s study examines the relationship between PFF’s cumulative by-team defensive stats^[2] and a team’s cumulative by-team defensive fantasy points. In other words, is there anything that a team’s performance in stopping the run, rushing the passer, defending the pass, and avoiding costly penalties can tell us about that team’s future ability to put up solid fantasy numbers? As I did with by-team offensive stats in the previous post, I will test the team defensive stats for both convergent and predictive validity.^[3]

Method & Results

Note about PFF’s Premium Stats: Pro Football Focus began their current method of statistical tracking back in 2008, which means we now have four seasons worth of fastidiously collected data at our disposal. As mentioned above, the current work will be looking specifically at cumulative team defensive stats across all four seasons. There are five categories of defensive stats in total: Overall, Run Defense, Pass Rush, Pass Coverage, and Penalty. What the stats intend to capture should be intuitive and self-explanatory, but I’ll refer you here for more precise details. Although the numerical range varies for the different stats, generally, more extreme negative numbers indicate the team performed poorly in that particular facet of the game while more extreme positive numbers indicate they performed well.

The approach I took to totaling a team’s defensive fantasy points was quite similar to my method for totaling offensive fantasy points. First, I divided the defensive points into the following categories: points from Tackles (1 point for solo tackles, 0.5 points for an assist), Sacks (2 points per sack), Air Pass Defense (1 point per pass defended, 2 points per interception), Total Pass Defense (Air Pass Defense + Sacks), and Total Defense (Tackles + Total Pass Defense).^{[4] [5] [6]} After choosing the categories, I simply summed the totals in each category for every defensive player on the team. So, for example, the 2011 St. Louis Rams Tackles total of 784.5 comes from James Laurinaitis’s 122 points + Quintin Mikell’s 83 points + … + Ronald Bartell’s 1 point.

For this first level of analysis, we have 128 cases (32 NFL teams × 4 seasons) of the 10 variables described above (the five PFF stats and the five categorized fantasy totals). To establish the convergent validity of the five PFF measures, I measured the level of correlation between each of the PFF stats and each of the categorical fantasy totals. Results are summarized in the table below.

Note about interpreting correlations: Correlation coefficients can range from -1.000 to 1.000. The closer the coefficient is to either of those values (i.e., the farther it is from zero), the stronger the linear relationship between the 2 variables. Positive coefficients indicate positive linear relationships—as one variable increases, so does the other. Conversely, negative coefficients indicate inverse relationships—as one variable increases, the other decreases. When reading these tables, a single asterisk next to a coefficient denotes a statistically significant relationship between the two variables referenced in the intersecting row and column (p < .05); very significant relationships are indicated by a second asterisk (p < .01)

The first thing to notice here is that none of PFF’s five defensive stats have a meaningful relationship with the aggregate number of fantasy points being scored by a team’s defensive players (i.e., Total Defense). I was initially quite surprised by this, but it becomes apparent why this is the case when we take a closer look at the correlations between the stats and the categorized fantasy totals. Specifically, there are strong negative relationships between all areas of defensive performance (except for Penalty) and the amount of fantasy points from Tackles, but these negative relationships are offset by nearly uniformly positive, significant relationships between the PFF stats and points accumulated through sacks and pass defense.

So we’re seeing two basic relationships here: (1) bad defense begets higher tackling totals, and (2) good defense produces opportunities for pass rushers and defensive backs to make plays in the passing game.^[7] Both of these relationships make a lot of intuitive sense—if a defense is bad, that means that team is likely out on the field longer, which means there are simply more tackles to be made. On the other hand, if a defense is really good, it’s likely repeatedly putting the opposing offense in risky and predictable passing situations (e.g., third and long) where the defense is prepared to capitalize on the offense’s disadvantage.

As we did in Part 1, to further investigate this causal relationship between defensive performances and fantasy points in respective defensive categories, we’ll test the predictive power of each of the stats. To put it broadly, we’re now answering the question, “How well do PFF stats from a given season predict fantasy totals the following season?” And, again, as in Part 1, this analysis now uses 96 cases (down from 128 in the previous analysis)—i.e., 32 teams and 3 season-to-season comparisons: 2008 to 2009, 2009 to 2010, and 2010 to 2011. Again, bivariate correlations between the five PFF predictors and the five categorized fantasy totals (across those 96 cases) are shown in the table below.