Above and Beyond
by Sam
Sept 15, 2023
Methodology
There are 2 key metrics for determining each team’s projected fantasy points per game:
Player Projections
Positional Tiers
Player Projections:
Last year, I put together these projections during the dog days of summer… so I was working off of each team’s projected keepers and used the only player projections that were available at the time - ESPN’s. Since these projections were done before the FCS draft, I also didn’t have any DSTs to build into the projections, so each team was based on each team’s contracted (or projected to be contracted) QB, RB1, RB2, WR1, WR2, WR3 and TE.
This year, I made 3 tweaks. First, I used Fantrax’s projections. I thought they performed well over the course of the ‘22-’23 season, and take into account the FCS’s unique scoring system (which ESPN did not). Second, I waited until after the FCS draft so that I could build the projections based on team’s projected starters as opposed to their projected keepers. Doing so gives a more accurate look given that each team is working with a full roster. This also allowed me to make the third change, which was to build DST into the projections.
Player Projections are the easy part because I’m not making custom projections… yet ;). The real fun comes next.
Positional Tiers:
Assigning players to tiers is not a new concept, though I feel like it has gained steam in the last few years. For this exercise, I used FantasyPros consensus rankings since they aggregate rankings across a variety of contributors. The use of tiers allows me to quantify what an individual player’s boom/bust potential is.
A quick summary of how I did this:
- Compared a player’s historical projected points to that player’s actual points
- Calculated the standard deviation for each tier within each position group
- Assigned that standard deviation number to each player’s current year projected points based on their current year positional tier
Here’s an example of what that looked like:
What you’ll quickly notice in the above chart is that the higher tier WRs score more points per game and have less variance. Makes sense, right? We expect Tier 1 receivers to put up big numbers with little variance in between.
But as you get further down the tiers, you see more variance being introduced. Tiers 3-5 specifically stand out to me. WRs in this range historically average between 11 and 13 points per game, but there is much wider variance with players in this group (between 5 and 6 points). Oftentimes you’ll find “sleepers” or potential “breakout” players in these tiers, and this is why. When l looked at year over year performance, player’s in Tiers 1 and 2 were either there the year prior, or moved up from one of these tiers. It’s not typical to see a player come onto the scene out of nowhere and move multiple tiers year over year.
After looking at the difference between player projections & actuals and variance of scoring within tiers, I then use current year player projections to estimate each team’s starting lineup and determine what their most-likely outcome is, as well as their projected ceiling and floor.
Accuracy
All of that is cool, right? But can they be trusted?
Well, through 1 season, I’m feeling pretty good.
What you’ll see below are my projections (black line) versus each team’s actuals (green line):
There’s a consistent gap between my projections and each team’s actuals; however, you’ll notice that the curve is almost exactly spot-on. The only real variance in terms of curve is BRX (projections & actuals moved opposite directions), DCA (which are spot on), and LRC & REN (which both hit a much higher curve than what I had accounted for).
With that said, you may recall that I didn’t factor DST into my projections. I went back and added each team’s primary starting DST’s actual FPTs/G from the ‘22-’23 season. It’s not perfect, because I don’t have those team’s last year projections saved, but there’s not a ton of variance in DST projections so I feel pretty confident about how this shakes out:
On average, these projections came within 6 points per game of the actuals. Not bad, especially considering the length of the season and amount of roster turnover each team typically sees!
Boom/Bust
Now that we’ve established the methodology and accuracy of these projections, let’s take a look at this year’s:
No surprise to see Tucson leading all teams. Reno closely trails for 2nd in the West and TCB rounds out the Top 3 in the West, though the gap with BSC and COL are small. DCA leads the way in the East, but it’s neck and neck with CHA and LRC.
A few things that immediately stand out: 1) we’ve all been talking about it, but this year does truly appear to have more parity than recent years, 2) the projections must know Caden is a new father, because Portland is not looking hot, and 3) some teams have significantly higher upside than others. Let’s take a look at who the boom/bust candidates are on a few of those teams:
Reno - 37.2 FPTs/G Swing between Projection and Ceiling/Floor
Reno already boasts a strong team, with every starter position projecting double digit scoring per game. There is some obvious upside with Watson (if he returns to form), Pollard (now owning the Cowboys backfield) and White (seemingly the only serviceable back on the Bucs roster), but the model really sees boom/bust potential with Nate’s WR corps.
While I would be surprised to see any of those guys hit their floor numbers, I wouldn’t be surprised if their points per game were in the ballpark. In the same vein, I wouldn’t be shocked to see them near the ceiling either. Each of Hop, Cooper and Deebo have proven to be alpha receivers in their past. Hopkins in a new situation, Amari with improved QB play and Deebo with health could see each of them hitting on those ceiling projections.
Colorado - 37.2 FPTs/G Swing between Projection and Ceiling/Floor
While Chris’ squad doesn’t have the same baseline as Nate’s, they do have the same upside/downside. Similarly to Reno, most of Colorado’s upside is coming from the WR group. Both Aiyuk and Kirk are projected at 10.3 FPTs/G, but could hit as high as 18.3 and 18.2, respectively (which means they could also hit as low as 2.4 and 2.3). But I specifically want to highlight the upside the Icecats have with QB and RB:
Daniel Jones was a bit of a fantasy revelation last year, and Chris got him for a bargain in the FCS Draft. If he can expound upon last year’s success with a year of Daboll’s system under his belt and an improved receiving option in Waller, I could certainly see him tapping into the upside number. I think Dalvin Cook’s ceiling & floor projections are spot on. We know there’s talent there, but an older back sharing a room with the young & upcoming Breece Hall could force him into a marginalized role that hangs around that 6.5 FPTs/G number. On the flip side, I could see Aaron Rodgers wanting that vet in the backfield, him getting a lot of snaps & touches and hitting that borderline RB2 ceiling projection.
Charleston - 34.6 FPTs/G Swing between Projection and Ceiling/Floor
Charleston is set up nicely with big projection expectations for Fields, Saquon and Chubb. He could see massive swings with his WR, each who has pretty significant question marks.
DJ Moore - we know the talent is there, but will this guy ever get a QB? The Bears are hoping so, but I have my doubts. I would be shocked if he came anywhere near the floor number, but I don’t see him sniffing the ceiling number.
Mike Williams & Tee Higgins - can these guys stay healthy? Both have elite quarterbacks and have shown elite talent, especially in going up to grab those deep play, jump balls. If they’re able to stay on the field, expect them to hit close to the ceiling projections.
A consistent theme you may have noticed from this brief exercise is that, while RBs have the highest point potential with our 0.5pt per carry scoring, WRs have the greatest upside (and downside) in terms of boom/bust potential.
If you don’t like where your squad is sitting, you better pick up the phone. Numbers don’t lie.
