Week 3 Rankings & Game Flowbotics
Whether you’ve won or lost your first two matchups of the fantasy football season, there’s one thing we can all be happy about after two weeks. Information! We finally have real data points on every team around the league. We’re working with small sample sizes, of course, and each game creates new questions, but answers to our preexisting questions are coming into focus. Chris Carson is the Seattle running back to own. Carson Palmer and Eli Manning might be washed up. The tight end position is still a touchdown-dependent and injury-plagued minefield. Tony Romo is incredible in the booth.
Statistical data sets will continue to grow and improve, as well. Last week, we had to manage without Football Outsiders’ metrics on offensive- and defensive-line play. Those numbers are now generated, and you better believe they’re included in this week’s Game Flowbotics spreadsheet:
If you’re new to my world of game flow analysis and how it informs my weekly rankings, check out last week’s article for some baseline information and a link to a detailed primer on Game Flowbotics.
My Week 3 rankings can be found at the bottom of this article, but for the space between here and there, I want to discuss the ideas of ceilings and floors for fantasy football players.
Floors and Ceilings as Ranges
A lot of fantasy owners subconsciously understand this, but it’s important to understand that ceiling and floor for fantasy players are not fixed values. They are inherently unknowable, so it’s impossible to put a pin in a specific fantasy point value and say something like, “34.8 points is Player X’s ceiling.” Ceiling and floor are smaller, finite ranges within the full range of outcomes for any given player.
Think about it abstractly. Yes, there is a true theoretical or effective fantasy ceiling that is fixed. It’s the number of points required to render all other scoring irrelevant in a fantasy matchup. If you play in a PPR format and Jack Doyle catches a 3.4-yard pass on every down of every Colts drive through an entire game, he’ll rack up so many fantasy points that it probably won’t matter who the opposing team starts. We know this outcome is unrealistic, though. Balanced scoring formats take away any possibility of a player reaching his true theoretical ceiling.
On the other hand, a slanted scoring setup can help certain players or positions get close to a theoretical limit. For example, if your league uses standard scoring, but makes the mistake of adding points per pass attempt and points per completion, quarterback scoring becomes the determining factor for most matchups. Even if your running backs, wide receivers, and tight ends have great production, it might not matter if your quarterback is Deshaun Watson while your opponent has Drew Brees. Even if his specific outcome is unknown, both the floor and ceiling offered by Brees are high enough to make your non-QB scoring trivial, but I’m getting sidetracked.
The overall point is that outcomes are unknowable, and a player’s ceiling is a range encompassing all high-end scores. Marcus Mariota hitting 32 points versus Marcus Mariota hitting 36 points are both ceiling-level values for him. Well, they are and they aren’t. Who’s to say Mariota couldn’t score 43 points? He’s Schrödinger’s TItan.
Because I can’t expect to hit theoretical ceilings for individual players, the rest of my roster still needs to work on some basic level. We’re playing with probabilities and there are countless moving parts as our ranges of outcomes square off against those of our opponents. Even if my ceiling hits aren’t offset by dud performances elsewhere in my lineup, my foes can always hit higher ceilings to defeat me. Furthermore, ceilings often overlap, like if I have Julio Jones while my opponent has Matt Ryan. In DFS, drafters might use some of the exact same players in competing lineups.
Even worse, our players’ individual ceilings can compete with each other. If my top two rushers in a redraft league are Todd Gurley and Carlos Hyde, it’s extremely unlikely both will hit their ceilings on Thursday night this week, but they can still be my best plays at running back. Depending on the rest of my lineup, it might not matter which of the two hits his ceiling, just as long as one of them does. Ultimately, if we accept that ceiling is unknowable, but still projectable to some extent, that means ceiling itself is a range (and the same can be applied to floor).
Applications in Setting Lineups
In practice, when deciding between players, try to determine what ranges of outcomes you need from each of your lineup spots. Remember that different players score points in different ways. Guys like DeSean Jackson, Ted Ginn Jr., and Kenny Stills are boom/bust options via big plays, while guys like Jarvis Landry, Golden Tate, and Marqise Lee will typically provide a more stable and narrow band of outcomes thanks to higher volumes of shorter routes.
The goal is to set objective expectations for your various players. To be clear, this is an extremely difficult task. Things get even more difficult when you start comparing your needs for a given matchup with the various expectations of your players. It’s a game of balancing theoretical probabilities. For each lineup decision, ask yourself, “What are the odds of overlap between my needs and my expectations?”
The more you iterate through potential lineups, you’ll probably start to realize that the core of your team is set in stone. Your opponent’s lineup will have a similar set of locked-in players, and the comparison of each team’s core should dictate who is favored at a base level. Once you determine if you’re a favorite or an underdog, you can start playing around at the fringes of your lineup to find edges. Depending on your role, you may need to take bigger chances with certain lineup spots and start a Kenny Stills type over a Marqise Lee type.
In the end, however, these ideas are built on the fact that we can’t know everything. Intuition and instinct can only take us so far, so systematic tools and shortcuts are pretty much required when forecasting fantasy. The Game Flowbotics page is one of my go-to tools because it quantifies matchups with statistics. Rankings are another shortcut to help with player comparisons, but they’re more subjective. They tie players to single points of value, despite the fact that those players have overlapping ranges of potential outcomes. Keep that in mind as you dive in below.