Value-based drafting is one idea applied consistently: a player is worth the points he adds over the player you would otherwise have started at his position. Raw projected totals are close to useless for cross-position comparison, because different positions score on completely different scales. Subtracting a baseline puts them on the same one.
The idea is borrowed from baseball analytics, where value over replacement was formalised as a way of pricing players against a freely available alternative rather than against zero. Fantasy football adopted it because fantasy football has the cleanest possible version of "freely available": the waiver wire.
The three steps
1. Find the baseline. Count how many players at each position are in starting lineups league-wide. Twelve teams starting two running backs is 24 backs. Add flex slots — those are shared, so allocate them. The first player past that count is your replacement level. The calculator on the homepage does this for any slot configuration.
2. Subtract. For each player, value over replacement is projected points minus the projection of the baseline player at his position.
VORP = player points − replacement-level points at that position
3. Rank by the result, not by the raw total. That ranking is your board. It will look wrong. It is supposed to look wrong — it is the correction to the intuition that made it look wrong.
The comparison this makes possible
Two players, different positions. Enter what you think each will score and what you think the replacement-level player at each position will score.
Where the baselines usually land
There is no universal answer, because the baseline is a function of your league's slot counts. But the structural pattern is stable across most conventional formats:
| Position | Typical starting demand | Effect on the baseline |
|---|---|---|
| Quarterback | One per team | Baseline sits shallow; the pool below it is still productive |
| Running back | Two per team, plus most flex use | Baseline sits deep; production falls away quickly past it |
| Receiver | Three per team, plus flex use | Deepest baseline of all, but the pool is deep too |
| Tight end | One per team | Shallow baseline; the drop-off shape varies enormously year to year |
The single-quarterback case is the clearest illustration of the method. Quarterbacks post the largest raw totals in most scoring systems, and they are routinely the worst use of an early pick, because only one starts per team. Push the baseline down to QB13 and the difference between an elite quarterback and a freely available one is often smaller than the difference between a good running back and a freely available one — even though the quarterback scores far more points in absolute terms.
Then add a superflex slot and the arithmetic inverts. That is not a change in opinion about quarterbacks. It is a change in the denominator.
Three caveats worth taking seriously
Baselines are not points. The method needs a projection of the replacement-level player's production, and that estimate is where the real uncertainty lives. Getting the baseline rank right and the baseline production badly wrong will produce confident nonsense.
Games missed are part of value. A player who produces at a high rate for two-thirds of a season is not worth two-thirds of a full season's value — he is worth that, minus whatever the replacement you were forced to start cost you. Season-long totals already handle this if your projections are season-long; per-game rates do not, and mixing the two silently breaks the comparison.
Draft position is auction price in disguise. In a snake draft, VORP tells you the ordering. In an auction, it tells you the ratios, which is far more actionable: a player with twice the VORP of another should cost roughly twice as much of the discretionary money in the room.
What it does not tell you
Value-based drafting is silent on risk. Two players with identical VORP can have wildly different distributions, and in a short season the distribution matters — more than most people assume. It is also silent on roster construction: it will happily recommend a roster you cannot legally start.
Treat it as a corrected ranking rather than an oracle. It removes a systematic error that costs almost everyone points. It does not replace judgement about who scores what, which remains the hard part and always will.