High-Variance and Low-Variance Players, Defined

Every player on a projection board carries a median and a band. The median is the number that gets read; the band is the number that decides drafts. Our boards express that band as a P10 and a P90 — the model's tenth- and ninetieth-percentile outcomes for the season — and the distance between those two numbers is where variance starts on our board, not where it ends.

Width on our board is mostly a statement about history. Each band is how far players in the same bracket at that position have actually finished from their projections in past seasons, and every band on the board runs wide. A player's own numbers decide only which bracket he lands in: the published board has one band shape for the top of each position and a much wider one below it, and the line between them is sharp.

That makes the band a starting point for risk, never the full account. Two players can share a median to the decimal and be entirely different draft objects, and at one position the PPR board prints them with the same band. Drafting the median treats them as interchangeable; drafting the risk treats them as two bets with two different jobs on a roster. Where that risk comes from, how to add it, and how to spend it is the rest of this piece.

Four Risks The Band Can't Tell Apart

Role dispersion is the first. A back with a locked three-down job has a snap share that moves a few points week to week; a committee back's share can swing twenty points on game scripts nobody can see in August. Our model never reads a snap or a depth chart. It projects both backs off what they scored and handled over the last two seasons, and if their projections land in the same bracket, so do their bands.

Missed games are the risk readers most often mistake for pessimism. A soft-tissue history reaches the model only as games missed, through a recency-weighted count of the last two seasons, and the model can't tell a hamstring from a benching; that count scales the whole projection. If that drops a player below the top of his position, he carries the much wider band printed there, which is the only route missed games take into a band. In a real season this width is asymmetric, the floor dropping while the ceiling holds, so the bottom of the band is where you add it.

Efficiency dependence is the third. Production built on volume (targets and carries) repeats year over year more reliably than production built on touchdown rate or yards per touch. The model acts on that in the median: half of a back's or tight end's rate, and 60% of a receiver's, is re-priced from his volume at the position's league-average efficiency, so a season built on an above-average touchdown rate projects lower. A touchdown-heavy season on an expensive price is one shape behind a fade call; the bust and fade argument gets its own page and its own screen.

Thin samples come last, and the model treats them differently than readers expect. Each season is pulled toward a bottom-quartile rate for the position by its own games, a season under four games is dropped when the other one qualifies, and a player with no NFL stats in the last two seasons starts at that rate outright, so one full season is pulled no harder than each of a veteran's two. A projection pulled low can fall below the top of the position, where the board prints its wider band, so a thin sample reaches the band only through the bracket. How that rule works and where it is weakest is documented on the model methodology page.

Scoring format moves real variance before a draft starts. Reception volume is among the stickiest stats in football, so full PPR steadies pass-catching backs and slot receivers, while standard scoring pushes those same players back onto touchdown rate. Our band carries the same proportions into every format and picks its bracket from the full-PPR total, so a pass-catching back keeps the band shape his PPR total earns in standard, and the widening there falls to you. Format math is settled on the PPR, half-PPR, and standard breakdown; assume it here.

Reading Band Width On An Invented Stretch Of Board

Take an invented stretch of receiver board — the numbers below are made up to show the shape, and no run of our model produced them. Four receivers sit next to each other, priced within a round of one another, projected in full-PPR points for the season.

On the board, all four print one band shape, because they sit at one position in one projection bracket. Receiver A projects a median of 198, with a P10 of 75 and a P90 of 297. Receiver B medians at 194, P10 74, P90 291. Receiver C medians at 191, P10 72, P90 287. Receiver D medians at 186, P10 70, P90 279.

Start with the medians. Top to bottom, the four are separated by 12 points across a 17-game season, about seven-tenths of a point per week. The width column adds nothing to that: every band is the same shape, so it descends exactly as smoothly as the medians do.

Now add what the board can't see. Say A and C hold locked roles with two full seasons behind them, and keep the board's shape for both. B's last season leaned on a touchdown rate his target volume never supported, so widen his band at both ends, to 50 and 315.

D took over a starting job halfway through last season, and the model averages the bench weeks where he saw the ball at all with his starter weeks. His ceiling runs past the board's and his floor below it: 45 to 320.

The ordering flips depending on which end of your band you read. Sorted by P10, the stretch is A at 75, C at 72, B at 50, D at 45, and the distance from best floor to worst is 30 points. Sorted by P90 it runs D at 320, B at 315, A at 297, C at 287, and the distance from best ceiling to worst is 33. Same four rows, three different drafts, and the board's own P10 and P90 columns simply repeat the median order.

When To Buy Width And When To Buy Floor

The rule is roster-relative. Width earns its price when a roster already clears the playoff line without it, since a median outcome that already gets you there makes the ninetieth percentile the only draw that changes the season. In that position, receivers B and D, as you read them above, are the only two rows worth the pick.

Narrow bands earn their price when a single collapse ends the year. Early rounds, shallow benches, and a thin wire all convert one P10 outcome into an unrecoverable start, and A and C are what survive that. The floor difference in the invented stretch runs 30 points, a bit under two a week, and it gets paid out precisely in the weeks a thin roster can least afford a zero.

Variance also stacks, and this is where drafters lose the thread. Six wide bands in a starting lineup do not average out across 17 games the way they average out across a thousand seasons; you get one draw, and the swing on a lineup of six coin flips is enormous in both directions. Two or three wide rows in a starting nine is a posture; six is a lottery ticket with an entry fee.

Payout structure moves the answer more than most drafters admit. Winner-take-all pays the ninetieth percentile and nothing else, which makes width close to free. A league paying the top three pays medians and floors, and the same wide row that was free becomes expensive.

Band width says less about tier breaks than it seems to. Neighbors in one bracket share a shape and overlap across most of it, so the break comes from the spacing between medians, an argument made in full on the tier-break page.

The bottom of the board is where width is cheapest. A wide P90 at pick 140 costs nothing a roster would otherwise start, which is the entire economic case for the archetype; the sleeper screen is this same width argument run against late ADP.

Where This Sits In Your Draft

Band width is the second pass. Sort by price, cut the board to the players your picks can actually reach, read the board's bands across whatever survives that cut, add the width only you can see, and let roster shape choose between the wide rows and the narrow ones. The P10 and P90 columns sit beside the median on the draft rankings board, and they move whenever the model is rerun.

One caution about the output itself. A wide band is the most comfortable thing a projection system can publish, because a band that covers everything is technically never wrong. Ours cannot widen a band just because a player is hard to call, since the shape follows the bracket his projection lands in, so the live question runs the other way: whether the shapes are wide enough.

That is a question the record answers. The season backtest's band coverage is on the methodology page, and the public accuracy ledger runs the same test on the weekly model's ranges every week — read the record, then decide how hard to lean on any of it, including this piece.