Value over replacement: why the highest-scoring player is not the best pick

A player is worth what he scores minus what you could have had for free. That one subtraction is the difference between a ranking list and a draft board.

Updated 2026-08-24

The problem with points

Rank every player by projected points and quarterbacks fill the top of your board. They throw for 4,000 yards and thirty touchdowns; nobody else comes close. Draft that way and you lose, every time. Everyone in fantasy football knows this and most explanations of why are hand-waving about "positional scarcity". Here is the actual arithmetic.

You are not choosing between a quarterback and nothing. You are choosing between a quarterback and the next quarterback you could still get. What a pick is worth is the gap between the player you take and the player who would have started for you anyway if you had spent the pick elsewhere. That second player is replacement level, and the gap is value over replacement.

Where replacement level sits

Replacement level is the projection of the last player at a position who would still be starting somewhere in your league. In a 12-team league that starts one quarterback, eleven other managers also need one, so the twelfth-best quarterback is available to somebody in every draft — and everyone above him is only worth the margin over that.

Flex spots complicate it, and they are handled by how flex spots actually get used rather than by rounding: a RB/WR/TE flex is filled by a running back about 45% of the time, a receiver 48%, a tight end 7%, so a single flex adds those fractions to each position's demand before the replacement rank is worked out. A superflex counts as 0.85 of a quarterback, because a small share of superflex slots get filled by somebody else.

In a 12-team league with half-PPR scoring and one flex, that puts replacement here:

Read the last two columns together. The best quarterback projects 371 points and the best running back 307 — the quarterback is ahead by 64. But the quarterback you could have had for free is worth 290 and the running back you could have had for free is worth 167, so after the subtraction the back is ahead by 59. Nothing about the projections changed. The comparison did.

It moves with your league

Replacement level is not a property of a player. It is a property of your settings, and it moves whenever they do. More teams means deeper starters, which means a worse replacement player, which means everyone above him gains value:

The same effect runs through every setting. Start three receivers instead of two and the receiver replacement drops a full tier. Start two quarterbacks and quarterback replacement falls off a cliff — which is the whole of why superflex leagues draft so differently. Add a point per catch and the projections themselves change before any of this arithmetic runs, which is the other big lever.

What the board looks like once you subtract

The top of a real 12-team half-PPR board, showing both numbers so you can see the reordering happen:

The three things people get wrong

• Treating it as a ranking to memorise. It is a calculation, and it has your league's settings as inputs. A VORP board built for a 10-team standard league is wrong for your 12-team full-PPR league in ways that cost real picks.
• Ignoring that it changes mid-draft. Once six quarterbacks are gone, the quarterback still available to you is worse than replacement level was at the start. Value at a position rises as it empties — this is what a positional run actually is, and why the board here re-ranks after every pick rather than sitting still.
• Applying it to kickers and defences. The maths works, but the gaps are so small and so unpredictable that the answer is always "take them last". VORP is most useful exactly where the spread between best and replacement is widest.