Three Sites, Three Numbers, Zero Answers
Why fantasy projections contradict each other — and what we do instead
Open three fantasy football sites on a Sunday morning and look up the same running back. One has him at 14.2. Another says 11.6. A third says 16.9. All three are labeled “projection.” All three are presented with the same clean confidence. And not one of them tells you which is right.
So what do you do? Average them? Pick the one that agrees with your gut? Go with whichever site you paid for?
That’s not analysis. That’s a coin flip wearing a spreadsheet.
The problem isn’t that these sites are lazy. Most of them employ smart, hardworking analysts. The problem is structural — it’s baked into how the industry builds numbers in the first place. Understanding why is the fastest way to understand what we do differently.
The committee problem
Most major fantasy sites run projections the same way: a staff of analysts each build their own numbers, usually for their own assigned slice of the player pool. Analyst A covers AFC North running backs. Analyst B handles NFC West receivers. Each one brings their own model, their own assumptions about game script, their own read on a coaching staff’s tendencies, and — inevitably — their own biases.
That produces a real problem the moment you try to compare two players. If Analyst A is systematically optimistic about volume and Analyst B is conservative, then a 15.0 from A and a 15.0 from B are not the same number. They’re two different measurements taken with two differently calibrated instruments. Stacking them into one ranking is like combining Celsius and Fahrenheit readings into a single list and sorting it.
Now scale that across sites. Every outlet has a different committee, a different house style, a different editorial incentive. Sites that also run content businesses have a quiet pull toward interesting takes, because “everyone knows he’s a WR2” doesn’t drive traffic. Sites with DFS affiliate deals have a nudge toward players who make for exciting lineups. None of this has to be conscious or cynical to distort the output. It just has to be human.
The result is the situation every fantasy manager knows: a wall of numbers that disagree, with no principled way to adjudicate between them.
Why consensus doesn’t rescue you
The industry’s answer to this is aggregation — take a dozen projection sources and average them into a consensus. It feels rigorous. It mostly isn’t.
Averaging works when your inputs are independent, unbiased estimates of the same quantity. Fantasy projections are neither. Analysts read each other. They pull from the same beat reporters, the same target-share charts, the same Wednesday practice reports. When one influential source moves a player, others drift toward it. What looks like twelve independent opinions converging on truth is often two or three original ideas echoed ten times.
And when the inputs genuinely do disagree, averaging destroys the most useful information in the set. A player projected 8.0 by half the sources and 20.0 by the other half averages to 14.0 — the exact same output as a player everyone agrees is a rock-solid 14.0. Those are wildly different players. One is a coin flip on a workload question. The other is a metronome. The consensus number erases the distinction precisely when you most need it.
The deeper flaw: a single number can’t describe a football player
Even a perfect projection, from a perfect analyst, with no bias at all, has a fatal limitation if it’s expressed as one number.
Fantasy scoring isn’t a normal distribution centered neatly on an expectation. A wide receiver projected for 12 points doesn’t score 12 points. He scores 3, or 6, or 9, or 27. The 12 is an average across outcomes that mostly never happen. Two players can share an identical projection while having completely different shapes: a possession receiver with a tight cluster of 9-to-15-point games, and a deep threat who posts a 4 most weeks and a 30 occasionally.
In a season-long league where you need a safe floor, those are opposite recommendations. In a large-field DFS tournament where you need to beat thousands of lineups, they’re also opposite recommendations — in the other direction. A single point estimate cannot tell you which player fits which spot, because the information you need was thrown away before the number reached your screen.
That’s what a linear projection is: a summary statistic presented as if it were a forecast, with the risk stripped out.
What we do instead
We start with the market, not with opinions
Every number on our platform begins with real-time sportsbook player prop odds — pulled from FanDuel, DraftKings, BetMGM, and Caesars.
Prop markets are the most efficient and liquid mechanism that exists for pricing individual player performance. Real money moves those lines. Sharp bettors attack anything mispriced, and books adjust within minutes. By the time a Sunday slate kicks off, the market has absorbed injury news, weather, snap-count reporting, and practice participation faster and more ruthlessly than any human staff can.
Starting there means we’re not asking one analyst what he thinks about a player’s workload. We’re starting from the same numbers the books are willing to take action on.
We simulate the season into existence
Those odds feed statistical models that run over 100,000 Monte Carlo simulations per slate. Each simulation plays out a different version of the games — different target distributions, different touchdown outcomes, different game scripts.
This is the step that recovers everything a point estimate throws away. Run a slate a hundred thousand times and you don’t get a player’s average. You get his entire range: how often he busts, how often he pays off, how often he wins you a tournament by himself.
We report probabilities, not point estimates
Instead of telling you a receiver “projects for 13.4,” we tell you his odds of clearing any given scoring threshold. Every player carries a unique probability curve.
This changes the question you’re asking. You stop asking “who has the higher number” and start asking the question that actually determines outcomes: what are the odds this player does what my roster needs him to do this week? Those are different questions, and only one of them has a useful answer.
We build lineups that account for risk
Our lineup builder weighs those probability curves rather than stacking the highest projected totals. The mathematically “optimal” lineup on paper — the one every projection-averaging tool converges on — is also the lineup thousands of other people are building from the same public numbers. Winning a large field requires a probability edge over that field, not agreement with it.
You set the risk tolerance. The math handles the construction.
We stay current
Projections refresh every 15 minutes, with round-the-clock player news from our data partners at SportsDataIO and RotoBaller keeping injuries, inactives, and depth-chart movement folded in continuously.
This matters more than it sounds. A projection published Tuesday and never touched again is a historical document by Sunday morning. Markets move all week. Ours move with them.
What you should actually expect
No bias. No hot takes. No gimmicks. No get-rich-quick promises.
We’re not going to tell you we’ve solved football, because nobody has. What we will tell you is exactly where every number comes from, exactly how much uncertainty sits behind it, and exactly what odds you’re accepting when you click a player into your lineup.
The industry standard is a number with a name attached and no explanation of how confident anyone is in it. We think you deserve the probability instead — and the ability to see, for every single player, the full range of what Sunday might actually look like.
Fantasy Football Dashboard promotes responsible gaming. Please play responsibly. Gambling problem? Call 1-800-GAMBLER.
