I’ve argued that we shouldn’t rely too heavily on baseline projections. The issue isn’t that they don’t have utility; rather, it’s that we should be willing to consider beyond a single potential outcome. That said, working to identify the most likely outcome is inarguably a productive exercise.
So let’s walk through how to do a baseline team projection, incorporating injury risk and the potential ramifications, using an actual example - the New Orleans Saints.
Working Back to Front
Let’s start by looking at a finished product. You’ve probably seen a team projection that looks something like this on Twitter or elsewhere.
Because this is just an example, I rounded off numbers where possible. Don’t be afraid to get more granular with your rates.
But overall, this is a pretty plausible projection for New Orleans in 2017. Let’s discuss the steps I took to build this and the assumptions underlying the final numbers.
Team Level Trends
The first step is always understanding volume and run/pass split in the offense.
Drew Brees has missed just 2 games since coming to New Orleans before the 2006 season, and only 1 was due to injury. His health has allowed the Saints to be consistently pass-heavy, as they’ve thrown for more than 650 pass attempts in 7 consecutive seasons. Here are their total plays, pass attempts and rush attempts over those seasons, measured as a percentage of league average.
These are remarkably consistent rates as far as these things go. For other eams, adjustments are often necessary to account for things like game script and play-caller or personnel changes. Leading teams run more plays and rush at a higher rate, so if you think a team played over its head in a prior season or are due to improve in 2017, you should be considering adjustments to their historical rates when projecting forward.
The Saints have been almost impervious to game script’s impacts on run/pass ratio, and have also kept the Payton/Brees duo in place for years. (Payton did miss the 2012 season as a result of Bountygate). Brees’s health is a lynchpin to keeping this equation consistent for 2017, but we have him as a moderate-to-low injury risk. So let’s project him to play all 16 games and keep the team-level numbers close to what we’ve seen in prior seasons.
Note that the pass and rush attempts don’t add up to total plays. That’s because of sacks, for which the league average has been around 35-40 over the past few years. (The Saints tend to be better than league average in this area.)
Though there are way more answers needed on the efficiency side of a projection, doling out volume is the most crucial aspect to projections. Efficiency is incredibly variable year to year, and thus it’s difficult (and perhaps careless) to project extremely high or low efficiency. This leads to regressing a lot of efficiency stats toward a mean, which in some cases might be a league average, but in others might be team-specific (as we’ll see below).
Let’s start allocating volume for the Saints on the rushing side. First of all, Brees is not a rushing QB. He has accounted for between 15 and 35 rush attempts in each of the past 10 seasons, and he’s been between 21 and 27 attempts 7 times in that span. Putting him down for 25 attempts drops us to 380 attempts for the rest of the offense.
Next we have to consider how touches might be split among the RBs. Is there a lead dog? Is it a committee?
For the Saints, it’s generally been multiple backs. This table admittedly ignores the impact of injuries, but shows how they’ve relied on three backs with different statistical profiles.
From right to left we have a bruiser (high carry-to-target ratio), a passing back (low carry-to-target ratio), and a back who does a little of both, and is generally the touch leader.
After filling the bruiser role early in his career, Mark Ingram has seamlessly taken over the “RB1” role since Pierre Thomas left New Orleans. Starting in 2014, he began seeing both more rushes and significantly more targets per game.
Even with the addition of Adrian Peterson, I’m projecting Ingram to play a similar role in 2017. Peterson fits as more of the between-the-tackles bruiser, though he’ll probably see higher-end attempt numbers relative to the above table. Rookie Alvin Kamara looks to be the future passing-down back.
But both Ingram and Peterson come in among the top 27 most likely RBs to be injured this year. Should one miss time, I’d expect the other to see a slight uptick in volume, but the Saints have also not hesitated to move a depth RB into the rotation in these situations. To account for this, I’ve left some volume on the table as it seems probable another back -- possibly Daniel Lasco -- is forced into some playing time.
This is where projections require some discretion, and also open up room for debate. My decision to give Ingram 175 carries, Peterson 150, and Kamara just 35 -- given it’s his rookie season and the status of the other two backs on the roster -- factors in both some injury risk as well as some potential for a volume boost should another back get hurt. Kamara is arguably undersold here given the potential for injury to either of the two backs ahead of him, but tradeoffs are necessary as there is only so much volume to go around.
On the passing side, Brees targeted the WR position significantly more in 2017 than in prior seasons, and the added targets came at the expense of TEs.
Meanwhile, RBs should be expected to see well over 20% of targets, and I have them at 25%, spread among the 3 backs discussed above, and roughly in line with the roles I outlined. I’ve left some targets on the table for Travaris Cadet, should he make the team and earn some of the passing-down work (he saw 54 targets last season).We shouldn’t expect Coby Fleener to command the type of volume Jimmy Graham did in this offense, but it’s fair to assume some of Brandin Cooks’ targets will shift to the TE position in 2017. The 135 targets I allocated for TEs in my projection equate to 20.5%, so I’m essentially splitting the difference between 2016 and prior rates. Fleener gets the lion’s share as he is a relatively safe TE on the injury front, and still has the contract to dictate an important role in the offense.
That leaves 54.5% of targets (360) for WRs, a downtick from 2016 but still a high number relative to historical rates. Let’s start with new addition Ted Ginn. Ginn will probably run a lot of the deeper routes Cooks used to account for, but he won’t assume the total volume Cooks did. I’ve projected him for 70 targets, more in line with the roles guys like Robert Meachem, Devery Henderson, and Kenny Stills played in this offense. Additionally, Cooks has seen between 6 and 8 carries per year over the past three seasons (and Stills and Meachem combined for 5 in 2013), which is why I assigned Ginn 5 rush attempts as well.
As far as #1 WR Michael Thomas is concerned, massive passing volume hasn’t meant huge target numbers in this offense historically. No Saints player has hit 150 targets in Brees’ 11-year run in The Big Easy, and just 6 times has a player eclipsed 130 targets (Marques Colston and Graham each did it 3 times). I’m projecting Thomas to take a step forward from the 121 he saw in 15 games last season and crest that 130-target threshold, but just barely.
Willie Snead is coming off seasons of 101 and 104 targets, both in 15 games. Essentially the new-age Lance Moore to Thomas’s Colston, he gets a slight bump forward in his third season to 110 targets. This accounts for what Snead has already been able to accomplish, as Moore only broke 110 once, in the 2008 season where Colston missed 5 games.
I’ve essentially projected each of the top three WRs to stay healthy, and they do have favorable injury outlooks for 2017. Still, it’s important to leave some volume on the table for additional receivers, as even in a mostly healthy 2016 for Saints’ WRs, both Thomas and Snead missed a game and Brandon Coleman and Tommylee Lewis combined for 49 targets.
As noted above, efficiency is a tricky thing to account for because it isn’t very stable year over year. The logical conclusion is we should regress efficiency heavily toward the league mean, but there are a couple situations to keep an eye out for.
Look for players who maintain efficiency rates consistently above or below expectation across multiple seasons. Dez Bryant’s and Eric Decker’s TD rates are prime examples.
Beyond that, it’s important to consider team rates. This is where New Orleans is a great example. Drew Brees has maintained an incredibly high completion percentage over the past few seasons. RBs in particular catch a high percentage of the targets they see in this offense (and they see a ton). Meanwhile, Michael Thomas showed how that can translate to WR with a 76% catch rate in 2016. I’ve regressed that pretty substantially in my above projection for 2017, but 69% is still well above the 75th percentile for league-wide WR1s (and is arguably still optimistic considering Thomas will surely see increased defensive attention).
Another good example in New Orleans is RB TD rates. For the roles Mark Ingram and Adrian Peterson will fill, we should expect a high number of TDs relative to their overall carry numbers, based on the previous performances of backs like Chris Ivory and Khiry Robinson in this offense.
For offenses that don’t have these established trends, consider the overall strength of the offense. You’ll be more likely to see below-average efficiency rates in generally inefficient offenses, and vice versa.
To that end, keep the QB in mind as you fill out receiving numbers. Your receiving numbers will ultimately need to add up to what you have projected for the passers. In this case, beyond Brees’ high completion percentage, he consistently throws for around 5,000 yards and 30+ TDs. Few passing offenses can hit these rates, though.
The last piece of advice I have for efficiency stats is consider the role of the player. In my above example, Ted Ginn is carrying a substantially lower catch rate than Brees’ average. That’s because I anticipate a high percentage of Ginn’s targets to come well downfield with a high depth of target, and those passes naturally carry a lower success rate.
The counter to that is I have Ginn projected for high yards per reception, which is another rate where it’s important to consider role and average depth of target. To get a better understanding of a receiver’s depth of target profile, I recommend checking airyards.com, an awesome free site put together by Josh Hermsmeyer.
Ultimately, efficiency rates turn projected volume into fantasy points, so they are extremely important to any projection. That said, they require the most discretion. You should be careful to avoid projecting outlier rates, but this is the area of a projection where you have the most leeway for personal preference. It’s also where the widest ranges of potential outcomes exist, which is another reminder not to rely too heavily on your finalized projection.
Once you’ve established individual efficiency rates, the projection might feel like it’s done. But there are a couple of important checks to do.
First, make sure the combined receiving numbers add up to a projection in line with what you’d expect from the QB. It will probably require some tweaking while you ask yourself which players stand to benefit or lose the most production if your QB numbers are unrealistic.
Similarly, be sure to cross-check your rushing totals, and see that they are in line with the offense as a whole. TDs are an easy one to let slip through the cracks, but if your combined numbers create an offense scoring far more or fewer TDs than should be expected for the offense as a whole, you have to ask yourself which players need adjustment. I have a high number of TDs projected in this example (50) because the Saints have scored between 48 and 55 in each of the past 5 seasons. Not every team has that potential, obviously.
Once you’ve ensured the overall team numbers are in line with what you’d expect, you’ve built out your team projection.
One down, 31 to go!