Glossary

The vocabulary AutoTiers uses, defined the way we actually mean it. Some of these are league-wide fantasy terms; others are specific to how our engine turns projections into tiers. Where a term drives a scoring rule, we link the guide that explains it in depth.

VBD (value-based drafting)
A method for comparing players across positions on a single scale. Instead of asking "how many points will this player score?", VBD asks "how many points does this player score above the player you could get for free at the same position?" A quarterback who scores 300 points is not necessarily more valuable than a running back who scores 220, because good quarterbacks are plentiful and good running backs are scarce. VBD makes that scarcity explicit by subtracting a position-specific baseline — the replacement level — from every player's projected score. The remainder is the number AutoTiers actually ranks and clusters on.
Replacement level
The projected output of the last startable player at a position — the quality you can expect to find on waivers or with a late pick, essentially for free. AutoTiers sets it per position by multiplying your league size by a position factor and reading the player at that rank: running backs and receivers anchor at 2.5 × league_size (about RB30 / WR30 in a 12-team league, reflecting how many start across flex spots), tight ends at 1.25 ×, and single-QB leagues at 0.67 × (about QB8 — the last quarterback you can reliably stream, not QB12). In superflex leagues the quarterback baseline drops much deeper. That player's score is the zero point; everyone else is measured against it. See the methodology page for the full table.
Adjusted score
A player's blended projection after every enabled rule has been applied. AutoTiers starts from a weighted blend of last season's production and current consensus projections, discounts the prior-year term if the player missed games, then multiplies in each rule's effect. The adjusted score is what feeds VBD and, ultimately, the tiers.
Target share
The percentage of a team's total pass targets that go to one player. A 25%+ target share is elite — roughly one of every four throws — and is one of the stickiest, most predictive signals a receiver or pass-catching back carries from year to year, because it reflects a role the offense has chosen to build around them. It drives the Target Share Premium rule.
Carry share
The percentage of a team's running-back carries that go to one back. It separates a true lead back from a member of a committee. Under 50% flags a committee (a penalty); under 30% flags a pure handcuff or change-of-pace back. See volume & workload.
Snap share (snap %)
The percentage of his team's offensive plays a player was on the field for. It is the broadest measure of a player's role: you cannot produce on plays you watch from the sideline. A snap share falling under 55% usually signals a shrinking role, which AutoTiers can penalize.
Red-zone looks
Carries plus targets inside the opponent's 20-yard line — the touches most likely to become touchdowns. Because touchdowns are worth six points and are hard to project directly, red-zone opportunity is a more stable leading indicator than red-zone scoring. A heavy red-zone role (25+ looks in a season) earns a boost; see opportunity & usage.
Opportunity score
A composite of a player's targets, carries, and red-zone looks that estimates how many fantasy points their usage should have produced, independent of whether the ball actually bounced their way. Comparing real production against opportunity is how AutoTiers spots regression candidates in both directions.
Touchdown regression
The tendency of touchdown totals to drift back toward what a player's opportunity implies. Touchdowns are famously noisy: a player who scores three or more times above what his red-zone volume predicts is unlikely to repeat it, and AutoTiers treats that as a sell signal.
Jenks natural breaks
A one-dimensional clustering algorithm that finds the "natural" gaps in a set of numbers — the cut points that group similar values together and put the biggest jumps between groups. AutoTiers uses it to draw tiers, because the whole point of a tier is that the players inside it are interchangeable and the drop to the next tier is real. We use Jenks up to 11 tiers; above that the score distribution is smooth enough that equal-size quantile splits are more honest than letting Jenks chase noise.
Quantile splits
Dividing ranked players into equal-size groups (e.g. twelve tiers of equal count). AutoTiers falls back to quantiles when you ask for a large number of tiers, where genuine value cliffs are too small for Jenks to separate from random variation.
Superflex
A league format with a flex spot that can start a quarterback, so most teams start two. This roughly doubles quarterback demand and pushes the replacement-level quarterback far deeper (to about QB18 in a 12-team league), which sharply raises what elite quarterbacks are worth in VBD terms. AutoTiers adjusts the quarterback baseline automatically when you set two QB starters.
ADP (average draft position)
Where a player is being drafted on average across public leagues. ADP is a measure of market opinion, not projected value, so AutoTiers never feeds it into a player's score. It is used only as a tiebreaker for ordering players who land in the same tier. The FAQ explains why.
PPR / half-PPR / standard
Scoring formats that award one point (PPR), half a point (half-PPR), or zero points (standard) per reception. The choice reshapes the entire board: full PPR lifts high-volume receivers and pass-catching backs, while standard scoring rewards touchdown-dependent runners. AutoTiers re-scores every player for the format you pick.
Prior-year games discount
A weighting that keeps a partial prior season from dragging a player down unfairly. Last year's point total silently bakes in zeros for missed games, so AutoTiers scales the prior-year term by games played, reaching full weight around a 14-game season. An injured star's four healthy games no longer read as a full, terrible year.
← Back to AutoTiers