NFL DFS Strategy: DvP Rankings Year-Over-Year Correlation
Editor's Note: This is a free preview of one of our premium NFL DFS strategy articles from last season covering evergreen DFS topics. To read more content like this and access our research tools join us at DailyRoto premium this fall.
Still a few sleeps away from Week 1 and as of this writing the NFL hype machine has already been cranking for a few weeks and about to get dialed up another notch between now and Sunday. The early weeks of NFL can be the best of times and the worst of times as we try to learn what we can from offseason news, personnel and coaching changes, and vanilla scheme preseason games and employ it in a better way than the rest of the DFS community. I still remember licking my chops heading into the 2014 season ready to stream QBs against the porous Dallas Cowboys secondary until an improved offensive line and focused running attack headlined by a career year from DeMarco Murray saw them improve from the WOAT to a middling unit versus the pass from a Fantasy perspective. In this article we'll be exploring the risks and benefits of leveraging prior years' defense vs. position (DvP) rankings in our early season matchup analysis.
- Gather Fantasy DvP rankings from prior NFL seasons
- Plot data to identify trends and correlations
- Determine how we should leverage our findings and explore outliers
I've linked to some tables with the raw rankings data at the end so you can dive in team by team as you see fit, but for my general approach what I've done is plot out the rankings data with different years on the X and Y axis – if the rankings held perfectly true we'd expect to see a linear trend up to the right with the #1 ranked team remaining #1 and #32 remaining #32, and if they were completely random we'd probably see something close to a horizontal trend line. To do this I gathered 2011 to 2015 DvP rankings and plotted each year's ranking on the x-axis with the prior year's ranking on the y-axis, allowing us to see how the rankings correlated.
In aggregate we do see a correlation between DvP on a year over year basis, but by viewing the changes on a scatterplot we're able to more effectively visualize what changes happen – the teams on the top left are defenses that went from poor to elite, and the bottom right went from elite to poor, while there are definitely clusters of teams making only marginal improvements. The big takeaway for me was that in aggregate it is still rare for teams to make extreme leaps but a lot more frequent that they can make enough adjustments to become a competitive unit, but I wanted to also look at positional impact.
Quarterbacks, Wide Receivers and Tight Ends
So I started off thinking I'd analyze these positions individually but when the results came in it made more sense to pull it into the same section. Quarterbacks ended up being slightly more correlated than our aggregate rankings leading me to believe that teams had a harder time adjusting to “the passing attack” until I took a closer look at Wide Receivers and Tight Ends. When looking at that 56% of teams move at least 5 spots in their QB rankings and 31% move more than 10 spots, but for wide receivers 73% move at least 5 spots in their QB rankings and 41% move more than 10 spots.
So if QB rankings are more predictive and consistent year over year but their complementary partners are not what could be causing that? One hypothesis is that the teams are still allowing the same performance but the targets are getting redistributed between tight ends and wide receivers showing that maybe those specific positions aren't as predictive and should be look at together. The other hypothesis ties back into some other industry articles I've read outlining that Vegas odds are less predictive for wide-receivers, and is that wide receivers can accrue a higher % of their fantasy points from catches and yardage, which normalizes more between the top and bottom NFL teams in a given year, causing more randomness in the rankings. Regardless, it seems like leaning heavily on wide receiver DvP from prior seasons may be a leak.
The Running Game
As demonstrated in this chart the rushing game does have positive correlation year over year but not as strong as either our overall correlations or our quarterback rankings. 59% of teams move at least 5 spots in the RB rankings with 35% of the teams jumping a full 10 spots or more.
So that is a lot of charts, what are our takeaways?
- There is correlation between teams DvP year over year but it is not as strong as you may have thought.
- Quarterback rankings are the most consistent followed by RBs. Wide receiver and tight end rankings from prior seasons are marginally predictable and shouldn't be relied on.
- It sounds obvious but I think in practice anyone can be guilty of relying on that nice green 31st and scared by the red 4th ranking next to a player's opponent on DraftKings. We should absolutely not be relying on prior seasons DvP as our sole basis for ranking a possible matchup.
With 31% of QB ranking, 35% of RB, and 41% of teams moving 10+ spots in rankings year over year there are going to be teams that make huge leaps and others that fall off. With many people focusing on what we know from last season, we may be able to gain an edge simply by admitting what we don't know, and can likely gain a larger edge by doing our diligence on which teams can make a leap or fall.
Outliers and things to look for
We can learn a lot from outliers by examining if they were truly random or if there is a signal beneath that leads to the trend. The 2014 Cowboys season tuned us into the aforementioned DeMarco Murray / clock control game that lead to Dallas giving up less Fantasy points through the air so one thing to look out for is teams that look like they may have more success running the ball. One team this could signal improvement for is the Tennessee Titans who finished 27th against QBs last year and invested in their running game both via the draft (Derrick Henry, plus 1st round picks in 3 of the last 4 seasons) and free agency (DeMarco Murray), so while PFF still rates the Titans line as one of the worst we know it has been an area they are focused on improving. If they are effective at improving their running game we should see a decrease in the fantasy points they are allowing.
Another area that can cause outlier performances is coaching and scheme changes, which can create a change in the pace that a team plays at. The pace, as Mike Leone has outlined in his article, can have as much of an impact on fantasy scoring as the players' skills themselves. One example of this is Chip Kelly's departure from Philadelphia and their 29th and 30th ranked fantasy defense the past 2 seasons. Not being an NFL scout I'm inclined to let the experts comment on how much of that is personnel driven but it stands to reason that an Eagles team lead by Carson Wentz may play slower this year impacting not just their numbers but their opponents. Similarly Chip Kelly moves to a 49ers team that had already showed signs of being a below average unit and may play at an elevated pace.
Are there any outliers you are planning to attack? Hit me up on twitter @drewby417.
Be the first to know
Want FREE Fantasy and Betting Advice and Savings Delivered to your Inbox? Sign up for our Newsletter.