Think you can beat the fantasy football consensus?

The fRank Open 2026 -- Rules

The market's board is not fixed before entries open. Consensus is
captured daily and each capture publishes a day later, so what you see
is always yesterday's board, never today's. When you enroll, your
ranking set starts from the newest board that has published; the board
you are actually scored against is a different, later one -- the last
capture taken at or before your entry deadline. By design, that capture
is taken within a day of your deadline, and nothing publishes until it
is a day old -- so it is not public yet when entries close. The market
submits its board on the same deadline you do.

Every QB, RB, WR, and TE is eligible -- rankable and scoreable, whether
or not the market ranked him. Ranking a player the market never ranked
is a real call: it counts toward your 100, and if that player
outproduces whoever the market held at that same slot, the slot pays
you. A pick at an unscored position is skipped, and your board closes
up: your next eligible pick inherits the slot.

You must rank at least 100 eligible players to enter. At every one of
those 100 slots, your pick is compared against the locked board's
occupant of the same slot. Each slot-pair earns a score from the real,
weekly, full-PPR value both players actually produced above a
published replacement level. Slot weight decreases linearly from the
top of the board to the bottom, with no cliff at the edge of your
ranked players -- an unranked player is scored as if placed 101st.

Matching the locked board, slot for slot, scores exactly zero. That is
not a penalty -- it is the definition of the baseline: you are being
compared against yourself. Match an older capture instead and you are
scored against the locked board like any other board -- not capped at
zero in either direction. A correct fade also scores: ranking a player
lower than the locked board and being right about it earns points, on
the same terms as ranking a player higher and being right does.

Every player's maximum score impact is bounded. No single slot, and no
single player's outcome, can contribute more than a fixed, published
ceiling to your total. A bounded contribution is not a zero
contribution -- it can still matter in a close race.

A scored week is one with a recorded stat row. (Published as
availability rule stat_row_weeks_active_v1.) Every other week costs
the player half his position's replacement level -- applied
identically to every player, never a private judgment call made after
the fact. Charging nothing would mean a one-game season could outrank
a full one, since sitting out would beat playing below your own
position's baseline; charging the full amount would assume you never
replaced an injured pick, which nobody does. Half is a judgment call,
published as one.

That charge is counted against 16 expected games, not the 17-week
graded schedule, so nobody pays for his bye. Week 18 is not scored:
it is the week clinched teams rest starters, and not a week fantasy
leagues play.

The value curve is derived separately for each position, from that
position's own startable weeks. One curve shared across positions
would price a quarterback's week against a pool of running back and
receiver weeks, and quarterbacks would score negative by construction
rather than by play.

Every scoring parameter is estimated from seasons before this one. The
method is public now, in full: the formula, the slot weights, the
scored depth, the availability rule, and the missed-week charge. The
fitted numbers behind the value curve and the replacement levels
publish when entries lock, and are frozen from that moment -- nothing
about them is refit after kickoff. Because they are fit only on
seasons already played, there is nothing in them that could respond to
how this one turns out.

The board you are scored against is not one of these parameters: it is
not fixed in advance either, and which capture it will be is not
decided until your deadline passes.

Boards are sealed from other entrants until lock, and FlexRank does not
read entrant boards before the entry deadline passes. After lock,
boards, slot-by-slot calls, and field consensus publish together.