Stop drafting points. Draft the gap: value over replacement, explained with the 2026 board.

Here is the trap every new fantasy football drafter falls into once: quarterbacks score the most points, so quarterbacks must be the best picks. Then they spend a second-round pick on a QB, watch their opponent grab one three rounds later who scores nearly as much, and lose the league at running back.

The number that explains the trap — and gets you out of it — is value over replacement: a player's projected points minus the points of a replacement-level starter at his position (roughly, the player you could still grab off the board late or stream from waivers).

Points are position-relative

On the 2026 Ball Ranks board, several quarterbacks project for more raw PPR points than the top running backs. But the board doesn't rank them first, because in a 12-team league the 12th quarterback is startable and freely available, while the 24th running back is a weekly liability. The gap between an elite RB and his replacement is enormous; the gap between an elite QB and QB12 is much smaller.

That's what the VAL column on the board measures. It's why Jahmyr Gibbs and Bijan Robinson sit at the top of the PPR board with values north of +150 points over a replacement running back, while quarterbacks with bigger raw point totals sit rounds later.

Three rules the number gives you

1. Compare VAL, not PTS, across positions. Within a position, points and value order players identically. Across positions, only value tells the truth.
2. Scarcity is a format question. In full PPR, pass-catching backs and target-hog receivers gain value; in standard scoring, the board reshuffles. This is why the Ball Ranks board recomputes value per format instead of publishing one list with a disclaimer.
3. The cliff matters more than the peak. The right question at every pick is "how bad is the alternative if I wait?" — that's replacement level doing the work. Positions with a gentle slope (QB, and in most years TE after the elite two or three) can wait; positions with a cliff cannot.

Where our model argues with the room

One honest wrinkle from this year's board: Model One consistently rates mid-round quarterbacks — the Jalen Hurts / Jared Goff / Bo Nix tier — higher than the rooms currently drafting them. Their projected points over a replacement QB are real (Hurts projects about +43 points over QB12 in PPR), and the room lets them slide 20 to 40 picks past where the model would take them. You don't need to reach for a QB early — the whole point of replacement math is that you usually shouldn't — but when that tier is still sitting there in the middle rounds, the discount is real and the model says take it.

The reverse is also on the board: several hype picks with thin or nonexistent NFL track records are being drafted rounds ahead of anything their history supports. A projection built on evidence will always be late to a genuine breakout — we accept that trade knowingly, and the ADP column is right there so you can see every disagreement and make your own call.

Use it live

The 2026 draft board has all of this per scoring format: projected points, value over replacement, position rank, and live ADP from real drafts, for about 680 players — with a full profile page for every one of them, down to his team's actual depth chart.

Draft the gap, not the total.