Since this is the first season my player projections are available in the Sharp Fantasy Football Draft Kit, I wanted to take a moment before we are at the heart of drafting season to provide a primer on how you can use the projections in unison with the player rankings, the incoming tiers (starting this week), and ADP.
Click here to get the Draft Kit for just $1!
What are Projections?
Projections are simple.
My season-long rankings and projections focus on the probable outcomes for a player based on top-down production on a per-play basis and projected game script.
Player production is based on that team volume. We can tweak volume and efficiency for a range of outcomes per player, but that is the simplest explanation of how the projection sauce is made.
Projections operate as more of checks and balances since we inherently know they will not be perfect. The modeling is operating under probability and median outcomes based on previous player performance.
Over the summer, we will get a ton of commentary from beat reporters placing counting stats on expected player performance.
This is not a not to dunk on beat reporters because we need them and many do actual reporting, but you need to place a huge grain of salt on all of those because many beat reporters lack an actual understanding of how many opportunities are available per team and how finite those opportunities are. Many do not have that macro perspective as a job requirement.
This has already happened this offseason. The Athletic had a Steelers reporter suggest that 100 catches was in the realm of possibilities for Pat Freiermuth this season.
While vague enough to have outs, the probability of Freiermuth catching 100 passes this season is very slim, and not only because only four tight ends have caught 100 passes in a season over the past decade.
Let’s use this example to take a top-down walkthrough of Freiermuth's projection and how that number is improbable.
Freiermuth has been targeted on 20.1% and 22.6% of his routes run to open his career.
Even if we gave Freiermuth a year three growth up to a 25.0% rate for 2023 (for context Travis Kelce was targeted on 25.3% of his routes), he would need to run around 600 pass routes at his career catch rate to reach 100 receptions. For additional context, Kelce led NFL tight ends last season with 600 routes while just two others (T.J. Hockenson and Evan Engram) even ran 500 pass routes.
Now, there are pros and cons to using projections based on their inability to project outlier production outside of those probable outcomes.
That Freiermuth example was an extreme way of looking at things. Even though that 100-catch season is an unlikely outcome, he can still project to be undervalued, and we can run into a career performance as part of the ride.
The pro side is where those checks and balances come in since a player that has an outlier season versus his career norm will inherently be priced up the following season.
There will always be a tax to pay on career seasons in fantasy football, but using objective projections can reveal a sobering expectation in the follow-up season.
This was applicable in the case of Deebo Samuel last season, who was coming off a 2021 campaign that looked nothing like his previous output from a usage and efficiency stance.
Samuel’s 2022 projections were not poor by any means, but they revealed that he was overvalued as a locked-in fantasy WR1 when he carried an average draft position of WR8 last offseason.
On the flip side, projection models are never going to highlight those outlier seasons that tilt seasonal outcomes.
No model was going to project Jamaal Williams to lead the NFL in rushing touchdowns.
No model was projecting Josh Jacobs to lead the league in rushing.
And that’s okay if you are using projections to still guide you into players that are undervalued and offer red flags for those who may be overvalued.
Going back to that 2021 season, my projections and rankings were higher for Samuel than for the rest of the field. They did not project the top-end outcome that we received, but because his projections showed a better asset than his market price that offseason, I ended up with a ton of Samuel on my rosters.
That is all I am looking to do at the baseline with projections. Use them as a North Star in hopes they lead to one of those spike seasons.
No model was projecting Geno Smith to throw 30 passing touchdowns, but Tyler Lockett was a player that showed up a far better fantasy asset last season than his ADP. We ranked and drafted him accordingly.
A player such as Courtland Sutton inversely showed up as someone to proceed with trepidation on at his cost last season, and we correctly navigated those waters.
We are not always as lucky in the outcomes, but you can take the projections right now and compare them to the current market to find players that stand out in both directions.
Players Who Do Not Project Well
While it is imperative to not take projections as gospel, it is also imperative to understand that certain players and situations are just rarely ever going to project well.
Rookies are a great example since there is no actual player usage or performance from an individual player perspective to draw from.
We can use historic baselines based on draft capital to diagnose a rough outline for the potential opportunity out of the box, but that is a roughshod projection at best.
This is why rookies often end up being some of the better values in fantasy drafts.
Fantasy gamers love the upside, but they also love comfort. They love to know what they are buying.
This leads us to the next area where projections will not shine enough light, which is ambiguous depth charts.
Muddy situations are never going to project well, and rankings and ADP for players in those ambiguous depth charts reflect as much. This is also why some of the best yearly values come from these situations.
This season, we have several ambiguous backfields in Miami, Chicago, and Philadelphia, and the Giants and Chiefs have discounted wide receiver rooms since nobody stands out through objective conviction.
Devon Achane and Roschon Johnson fit the criteria of both. Jaxon Smith-Njigba is close since we have not seen the Seattle wide receiver run three deep, but we do know that both DK Metcalf and Tyler Lockett are established.
Projections Do Not Equal Player Rankings
You may have already checked out the projections and compared them to the player rankings and are wondering why there is not a 100% overlap between both.
Part of that ties into what we have already covered. You should always be thinking about the range of outcomes for players since projections are homed in on median and probable outcomes.
I can think of several examples of why my 2023 player rankings do not fully align with the projections, but I will use Chris Godwin as an example.
The projections have Godwin as the WR17 in full-PPR scoring. I have him as WR24. ADP has him at WR28.
I want to be above the field because Godwin objectively shows up as undervalued. There appears to be too much negative weight placed on the unknown floor for the Tampa Bay offense post-Tom Brady.
But I also do not want to be overly bullish on Godwin because of the company he keeps in rankings and ADP. While Godwin’s median numbers look undervalued, the upside scenario for players orbiting similar draft capital is higher-end outcomes, even if they are less probable.
Season Long Stats Are Not a Lie, but They Are Not Exactly Telling the Truth
Even if Godwin does hit his projection and appears to be someone who outproduced his ADP, how much value will he provide?
This happens every summer as fantasy football analysis takes a holistic look backward at previous production and upcoming projections for the current season. My hands are not completely clean in this regard as well.
But one thing I have always done is stress that season-long fantasy output rarely matters in the context it is often presented.
On a yearly basis, the number of players at each position providing elite weekly production for 17 weeks can be counted on one hand. The rest just survive, provide enough spike weeks, and we hope to be in on those players when their pockets of production hit.
Last season, Joe Mixon had 270 touches for 1,255 yards and nine touchdowns.
If we projected those totals for Mixon on the surface, that has the appeal of a solid season, right?
He was the RB10 in full-PPR scoring for the season.
Ask anyone who rostered Mixon last season if being the 10th-best running back in overall production really impacted their fantasy season.
Cam Akers was the RB35 last season. On the surface, he appears as a colossal disappointment compared to his draft capital, which was RB17.
But Akers impacted fantasy championships to a greater degree given when his best stretch of the season came.
While those projections give us a range of season-long production and have implications for listing players in a linear format (rankings), the one thing that is missing is that even when those full-season numbers are accurate, they are failing to capture the overall weekly impact and pockets of production that are relevant to our weekly game of fantasy football.
That is an anecdotal example to drive home a point, but outside of a few players each season at the top of their positions, you should not overly concern yourself with end of season standings or use that as any sort of litmus test for diagnosing true value.
So, in using Godwin’s projection, what does ending the season at WR17 really say if we get it?
Is he the type of WR17 that provides baseline-level production but never misses a game and accrues counting stats that are greater than his weekly influence on your lineup?
Is the type of WR17 that is volatile but has enough week-winning weeks to alter your season?
My thoughts are that in most outcomes he ends up being the former, which is why despite Godwin looking undervalued based on projections, I would still rather take swings at that projection cost on a player with more perceived upside in turning weeks such as a Christian Watson and then circle back to Godwin when his price point aligns more with that median outcome.
Consistency and its value has been an ongoing dialogue for fantasy gamers. As mentioned, there are very few players who are consistently high scoring on a weekly level per season, and fantasy scoring on a player level is not a bell curve. It is carried by spike moments among those players.
At the end of the day, year-over-year consistency is unstable. In-season variance for players can be alleviated across a full lineup.
When we lose, we are inclined to designate fault to the low individual scores not matching their per-game output, but the answer is almost always a lack of peak performances across the board in our lineups. Inversely, when we lose, our opponents likely stacked multiple peak performances.
Pair that with the knowledge that the average fantasy matchup outcome is not as tight to begin with as we believe as highlighted by Adam Harstad in this golden oldie on the subject, and what I am really saying here is that low weeks do not lose you games. High points win weeks for you, which is a tough hurdle for gamers to accept.
Synching up With ADP
While a player like Godwin may not be a week-altering player in the majority of outcomes, I still do want to follow through on my objective data suggesting that he is underpriced. We already do know he is a good enough player on his own merit to turn in something like Lockett did a year ago.
The good news here is that while Godwin’s projection may have him at WR17, you do not have to select him as the WR17. This is where the market helps you out. You can select him without reaching.
There are a few other players that fit this same criteria that I walked through using Godwin as an example.
His teammate Rachaad White, James Conner, Keenan Allen, and Brandin Cooks are all players that project higher than their current cost. They come with enough red flags through age, performance, or team environment that move them down in my rankings below that projection, but they also have a cost below those rankings where you can still take swings on them turning in an outcome that offers actual value.
The internet loves lists, and I am aware that most of you are here just for my answers to the test.
Still, I hope that if there is one takeaway here, it is that gamers do not think of the game so linearly as the way they are presented through player projections, rankings, and ADP.
