How a
player gets
a score
None of this predicts the future. What it does is measure the things that have historically carried over from one season to the next, and ignore the things that don't.
Written for three people who mostly just want to win an argument about a running back.
The one idea underneath all of it
Fantasy points come from opportunity far more than from talent. A mediocre running back who gets the ball twenty times a week will out-score a brilliant one who gets it six times. So almost every useful metric is really asking the same question: how much work is this player actually going to get?
Efficiency tells you how good a player looked last year. Opportunity tells you what he's likely to do next year. When the two disagree, back opportunity.
This isn't a hunch. Analysis of year-over-year correlations by SumerSports found usage statistics to be the stickiest measures in football — the ones that repeat. Rushing efficiency, by contrast, barely carries over at all. A back who averaged 5.2 yards a carry is not meaningfully more likely to do it again than one who averaged 4.1.
What each group of bars means
Opportunity
The most important group, and the reason the bars change depending on position.
For receivers and tight ends the headline number is target share — the percentage of his team's throws that come his way. Research collected by the Harvard Sports Analysis Collective puts it at explaining roughly 60 to 70 percent of the variance in a receiver's per-game fantasy points, and its year-to-year correlation sits around 0.70. Nothing else in football is that reliable. As a rule of thumb, a share above 25 percent is territory where genuinely elite seasons happen.
For running backs it's opportunity share: his cut of the team's carries and targets combined. PlayerProfiler's work on weighted opportunities found it the strongest available predictor of running back scoring, and it's why the app also shows how often his team runs the ball at all. A great back in a pass-heavy offence is a worse asset than an average one in a run-heavy offence.
Snap share backs both of these up. It's simply the percentage of his team's offensive plays he's on the field for. Below about 60 percent usually means he's sharing the job with someone, which caps his ceiling regardless of how good he is.
Availability
The coaching cliché is that availability is the best ability, and for fantasy it's close to literally true — a player scores nothing in the weeks he doesn't play. This group holds his current injury flag, how many games he managed last season, and how settled his role is.
Role certainty is the shakiest number in the app. It comes from the published depth chart, and depth charts move right through preseason. Re-run the data the day before you draft.
Form
Production is last season's points per game, ranked against others at the same position. Consistency measures how much he bounced around week to week — two players can average fourteen points while one delivers it every Sunday and the other alternates between two and thirty. For a weekly lineup you want the first one. The calculation is the spread of his weekly scores relative to his average, inverted so that steadier is higher.
How he finished compares his last five games to his season average, which is your "how did it end" question. Treat it gently: a player whose team was already eliminated may have been rested rather than declining.
Context
Everything happening around the player. How much his offence scored, how hard his schedule looks, and whether he's moved clubs. A move gets a small penalty rather than a large one — new teams sometimes mean a bigger role, but they always mean less certainty about what that role is.
How the composite is built
Six numbers go in, one comes out. Every input is first converted to a 0–100 percentile within the player’s own position — so a running back is only ever compared to other running backs. A score of 80 for production means he out-scored 80% of the backs in the pool.
| Input | Where the 0–100 comes from | Weight |
|---|---|---|
| Production | Percentile of last season’s fantasy points per game | 35% |
| Injury outlook | 100 if no flag · 70 questionable · 40 doubtful · 10 out or IR | 15% |
| Role certainty | 100 if first on the depth chart · 45 if second · 15 otherwise | 15% |
| Consistency | Percentile of week-to-week spread, inverted so steadier scores higher | 15% |
| Durability | Percentile of games played out of 18 | 10% |
| Finish | 50, plus 5 points per point his last five games beat his average by | 10% |
Multiply each by its weight, add them up, then apply two adjustments:
- Running backs only: add
(team run rate − 42) × 0.4. A back whose offence runs 50% of the time gains about 3 points; one at 35% loses about 3. Forty-two percent is roughly the league average. - Anyone who changed clubs: subtract 3. A deliberately small haircut — a move can mean a bigger role, but it always means less certainty.
The result is clamped to 0–100 and rounded.
A worked example
Say a running back scored well last year, played every game, sits first on the depth chart, was a bit streaky, faded slightly late, and plays for a run-heavy team he joined this offseason:
| Input | Score | × weight | Contributes |
|---|---|---|---|
| Production | 88 | 0.35 | 30.8 |
| Injury outlook | 100 | 0.15 | 15.0 |
| Role certainty | 100 | 0.15 | 15.0 |
| Consistency | 45 | 0.15 | 6.8 |
| Durability | 95 | 0.10 | 9.5 |
| Finish | 44 | 0.10 | 4.4 |
Subtotal 81.5 · run-rate bonus at 48% +2.4 · changed clubs −3.0 · composite 80.9 → 81
Notice what that shows: his consistency and late fade cost him roughly nine points between them, and joining a new team costs about the same as a decent run-rate bonus gains. If you disagree with those trade-offs, the weights live in one visible object at the top of the data script — change a number, push, and every score recalculates within a minute.
When a score is missing
A player with no prior-season data gets no composite at all, and shows a dash. That covers rookies, anyone who didn’t play, kickers, team defences, and the occasional player whose name didn’t match our data source. It is deliberate: a fabricated score would be worse than an honest gap.
What it cannot see
This is the part worth reading before anyone treats a score as gospel.
- Rookies. No prior season means no production, consistency or durability figures, so their composite collapses to near nothing. Rank them off the media consensus instead. The app flags them.
- Coaching and scheme changes. A new offensive coordinator can transform a player's value overnight and none of this notices.
- Anything from training camp. Beat reporters watching practice every day know things the box score won't show for months.
- Contracts, holdouts, suspensions — unless they've already reached the injury or status field.
- Touchdown luck. A player who scored on an unusually high share of his touches last year will probably regress, and raw points per game can't tell the difference.
- Strength of schedule. Working this out needs opponent defensive rankings, which our data source doesn't carry. That bar stays blank rather than guessing.
- Red zone targets for receivers. The data has red zone carries for backs and red zone attempts for quarterbacks, but no equivalent for pass-catchers.
How to actually use it on the night
Don't draft down the composite column. The scores and the media rank are shown side by side deliberately, and the gap between them is the useful signal.
When the model rates someone far above his media rank, it usually means the underlying volume is strong and the market hasn't caught up — worth a look. When it rates someone far below, it usually means his production came from somewhere fragile: touchdowns, a short hot streak, or a role he might not keep.
Either way the score isn't the answer. It's the opening line of an argument, and the argument is the fun bit.
Two practical things the numbers won't shout at you: don't stack players who share a bye week, and don't take two players from the same offence expecting both to eat — they're competing for the same targets.
Glossary
- PPR
- Points per reception. Every catch is worth a point on its own, before any yardage. It pushes receivers and pass-catching running backs up the board relative to pure runners.
- Target share
- The percentage of his team's pass attempts thrown at one player. The most reliable predictor in fantasy football.
- Snap share
- The percentage of his team's offensive plays he's on the field for. Establishes whether he has a full-time job.
- Opportunity share
- A running back's combined cut of his team's carries and targets. The running back equivalent of target share.
- ADP
- Average draft position — where this player typically gets taken across thousands of drafts. Compare it to where you'd have to pick him to see whether you can afford to wait a round.
- Bye week
- The one week each team doesn't play. Those players score zero. Two starters sharing a bye means a hole in your lineup you can see coming from August.
- Flex
- A lineup slot that accepts a running back, receiver or tight end. Flexibility to start whoever looks best that week.
- Committee
- When two or more players split one job, usually in the backfield. Bad for fantasy — it caps both of them.
- Regression
- The tendency for unusual results to drift back toward normal. Applies hardest to touchdown rates.
Sources
Usage and stickiness research from SumerSports. Target share variance figures via the Harvard Sports Analysis Collective, as summarised by Fantasy Analytics Authority. Weighted opportunity work from PlayerProfiler and Josh Hermsmeyer's WOPR. Player data and season statistics from the Sleeper API. Consensus rankings from Rotoworld.