We’re about a month out from the NFL draft. Most fantasy eyes are on what teams will do in upgrading their offensive skill positions or even their offensive lines. But what about the effect on fantasy team defenses of drafting players on that side of the ball?
Regular readers will know I’ve written several articles on fantasy team defenses this off-season:
- Team D/ST Scoring
- Fantasy DST Scoring: Does Base Alignment Matter?
- Fantasy Defense Scoring: Coaching Changes
- Fantasy Defense Scoring: The Impact of Changing Head Coaches
- Fantasy Defense Scoring: The History of This Year’s Coaches
A quick review of my methods: I’m studying from 2002 to the present and only looking at Team Defensive (Tm D) fantasy scoring, NOT defense AND special teams (D/ST) scoring. Here’s the scoring system I’m using:
- Interceptions = 2 FP
- Fumble Recoveries by the Defense = 2 FP
- Sacks = 1 FP
- Safeties = 2 FP
- Interception and Fumble Recovery Returns TDs = 6 FP
- 0 Points Allowed (PA) = 12 FP
- 2-7 PA = 8 FP
- 8-12 PA = 4 FP
- 13-17 PA = 2 FP (note NFFC scoring for PA may change for 2019)
All data in this study come from profootballreference.com’s (PFR) invaluable Play Index and Draft Finder.
I gathered up all the draft picks made by NFL teams following the 2002 season (i.e., the 2003 draft through last year’s picks) and then examined the change to the FP scored by the defenses teams making those selections.
In weighing the picks, I used two methods. One is the familiar Jimmy Johnson (JJ) draft chart, the other is Chase Stuart’s draft value chart. Those charts can be seen on Chase’s blog.1 You can follow the links there to learn how those values were developed. But basically, both methods put a premium on high picks and assign fewer points to lower picks:2
The Jimmy Johnson chart puts more value on the very highest picks compared to later first round and especially subsequent round picks. Chase’s chart drops off more gradually: his 1.10 has a value of 58% of the 1.01; Johnson’s 1.10 is just worth 43% of the top pick. And Chase’s 3.01 is 23% of the overall #1 while Johnson put it at 9%. It’s worth noting that Chase’s chart looks at the value returned by players over the first five years of their career while Johnson’s was developed more for trade purposes. But Chase has also shown that the JJ chart is more aligned with rookie performance, that is, if you are drafting for immediate help, the Johnson chart might better reflect the value of the picks while Chase’s chart assigns more long-range values.
I summed the value of the defensive picks made by each team in the last 18 drafts using both charts. Here’s the range of how teams have spent their draft capital on defense since 2002:
The 2017 Browns topped both systems in emphasizing defense in the draft (the Max category), using the #1 overall pick on Myles Garrett, another first rounder on Jabrill Peppers, plus a 3rd, 4th, and 6th on defense (and those were all high picks within those rounds).
Three teams had no picks on defense: Steve Spurrier’s 2003 Washington team, Rex Ryan’s 2009 Jets, and the Lovie Smith’s 2014 Tampa Bay. The Washington and Jets teams had only three picks each (that was the Mark Sanchez draft for the Jets). The 2014 Bucs had six picks, all on offense, headlined by Mike Evans.
To put the average (Avg) draft into perspective, a team picking 16th in each round would have about 43/1743 (Chase/JJ in draft capital) assuming it had its seven picks and no others. Around half of that would be used on defense (maybe a little more since defensive players have a little more value on special teams coverage). That’s about how the average draft has played out.
I ran regressions on the amount of draft capital used on defense (input) compared to the change (output) in Tm D FP from the year before the draft to the season after. For both draft pick value systems, there was little correlation to fantasy defense improvement (0.15/0.18, Chase/JJ). And that meant little of the change in Tm D FP can be explained by the use of draft picks on defense (R-squared values of 0.02/0.03, both statistically significant).
That was discouraging if you’re looking for a magic bullet to find this year’s breakout fantasy defense – but not surprising.
I thought maybe the numbers would be different if I looked at only those teams that heavily invested in D (i.e. more than one standard deviation above average in draft capital).
I divided the investment of draft capital on defensive players into five categories,3 then looked at how their “spending” affected their Tm D FP, first using Chase’s system:
The Previous Year FP is the average fantasy points scored by the TM D the year before the draft; the Next Year FP is that Tm D’s FP the season immediately after the draft.4 The FP% compares the FP scoring to the average defense in that year:5 so for the High investment category, those defenses scored 89% of the average Tm D in the previous year. The Change in FP is how much the teams in that category improved (positive numbers) or declined (negative numbers) after the draft. The final column is just the number of teams in each category; each is a decent-sized sample.
Not surprisingly, the teams that emphasized defense the most in their drafts were the worst fantasy defenses in the previous year. And those defenses were pretty close to average in the following season (98% of the typical fantasy D). So Tm D’s did seem to benefit from investing draft capital in defensive players. But remember a 10.4 FP improvement is not a huge number, and an “average” fantasy D is D16 or D17, which is not a starting defense in most fantasy leagues.
Here’s the same chart for the Jimmy Johnson values:
The results are similar, although the “High” teams were a little poorer quality and improved a lot more. Note the different draft value systems result in different categorization of the team drafts – there were 89 “High” defensive investments in Chase’s system, vs. 85 in Johnson’s.
I decided to dig one more level into the data. What if I broke the previous year defenses into quartiles (best 1-8, then #9 through #16, etc., similar to what I did at the end of this article):
Looking at Chase’s values, it looks to me like regression to the mean is as much a factor as drafting on changes in Tm D FP. Teams in the Bottom Quartile got a lot better whether they invested heavily in defensive picks or not. Very good defenses had a big drop in FP whatever they did in the draft.
But the Jimmy Johnson chart was a bit more interesting:
The average 2nd Quartile defense that spent heavily on drafting for defense (JJ system) typically bucked the trend of regressing to the mean. Remember the Johnson values tend to heavily weight immediate impact drafted players vs. the Chase Stuart system.
There are a few other blips in the data, and these sample sizes (here are only 12 teams in the 2nd Quartile/High group in that chart) are not necessarily meaningful. Personally, I think Tm D scoring is too volatile and NFL defenses too complex to be immediately affected by a few high-value rookies. But if you’re throwing darts at Tm D’s anyhow, check back after the draft and let’s see if any of last year’s 2nd-tier defenses spend a lot of Jimmy Johnson value on defense.
1I’m using the first of the two charts on the linked page.
2The extracted chart over-simplifies the draft by ignoring supplemental picks, for example, I used pick #97 for the 4.01 values. But most years, the actual 4.01 pick will be lower than #97 – in 2018, the first pick of the 4th round was actually #101 overall.
3The High and Low Categories are +/-1 standard deviation above the average amount of draft capital used on defense. The Average group is within +/- 0.5 standard deviations from average- roughly the middle group in defensive “spending.” The Above and Below Average categories then fall between High/Low and Average.
4So the FP in the 2018 season for the 2018 draft.
5Those averages and percentages are for the season in question, not the overall 17-year average.
