Most of Your Fantasy Draft Happens in the Fog

Seven spots.

This debate can involve podcasts, spreadsheets, YouTube clips, screenshots from beat reporters, and eventually a group text in which somebody becomes weirdly defensive about a backup running back.

Meanwhile, the people whose rankings we are arguing about may have that player 40 picks apart.

This is something I don't think we acknowledge enough.

The beginning of a fantasy draft looks precise because it actually is.

Then the precision quietly leaves the building.

Using the AutoDraftYears 2026 consensus board, the median disagreement between ranking sources is only 3.4 spots inside the first 25 picks.

By picks 101–150, it is 18.2.

After pick 150?

32.2 spots.

In other words, most of the draft takes place after the experts have stopped agreeing on what the draft board is supposed to look like.

Which seems important.

The first two rounds are basically a peace treaty

Here is how median source disagreement changes as you move through the 2026 board:

Median standard deviation of player rankings across contributing sources. AutoDraftYears 2026 consensus board, skill positions only.

That is a remarkably orderly progression toward disorder.

Early in drafts, everybody more or less knows who the expensive players are.

Bijan Robinson has a source spread of just 0.2 ranking places. Ja'Marr Chase is at 0.5. Jahmyr Gibbs, 0.6. Jaxon Smith-Njigba and Puka Nacua, 0.7.

Nobody is arriving at the draft with Bijan ranked 47th because they saw something interesting on film.

The market has reached a verdict.

You may disagree with the exact order, but everyone is shopping in roughly the same aisle.

Then you keep drafting.

And the aisles disappear.

After pick 100, everybody has apparently gone to a different Costco

Consider some of the most disputed players inside the top 150.

Sam Darnold has a ranking spread of 45.8.

Ricky Pearsall: 45.5.

Cam Ward: 41.1.

C.J. Stroud: 34.0.

Deebo Samuel: 32.7.

These aren't minor disagreements.

A 45-pick spread is roughly four rounds in a 12-team league.

One source is essentially saying:

We should probably draft this guy.

Another is responding:

Which guy?

That should change how we think about a consensus ranking.

Consensus feels authoritative because it produces a number.

Player X: 112.

Very official.

But 112 can mean two completely different things.

It can mean five sources independently landed somewhere around 112.

Or it can mean one source thinks 90, another thinks 135, and the spreadsheet has kindly averaged the disagreement into something that looks like certainty.

Those are not the same thing.

Standard deviation of the player's rank across contributing ranking sources.

This is where draft preparation should change

We spend months preparing for fantasy drafts, which is already objectively silly.

I have accepted this about myself.

But if we're going to do it, we might as well allocate our neurosis efficiently.

You probably do not need six hours of research to decide whether Ja'Marr Chase belongs at pick three or pick five.

The entire fantasy market has investigated Ja'Marr Chase.

The case file is approximately the size of the federal budget.

Your independent advantage there is likely small.

Compare that with a player ranked around 120 where one reputable source has him at 90 and another at 145.

That is where I want to spend my time.

Why are they disagreeing?

Is one source projecting a different role?

Is there uncertainty around quarterback play?

A new coaching staff?

An injury?

A rookie?

A depth-chart battle?

Or are we simply dealing with the normal late-draft chaos where everyone knows less and pretends otherwise?

The wider the disagreement, the more useful the question becomes.

Not necessarily the player.

That distinction matters.

Disagreement does not mean upside

This is where I need to stop us from doing the very fantasy-football thing we're naturally going to do.

A wide ranking spread does not automatically make someone interesting.

It definitely does not make him a sleeper.

In a companion AutoDraftYears analysis, controlling for draft price showed that expert disagreement itself did not predict better or worse fantasy outcomes. A wide spread tells us the market is uncertain. It does not tell us which side is right.

This is an important buzzkill.

Sam Darnold having a 45.8-place spread does not mean Sam Darnold is mispriced by 45 spots.

It means people have extremely different ideas about Sam Darnold.

Your job is to figure out why.

Think of disagreement as a blinking light on the dashboard.

It means:

Look here.

Not:

Draft this person immediately.

There is also a mechanical limitation worth acknowledging. Absolute ranking spreads naturally have more room to expand later in the draft because the rank numbers themselves get larger. The 32.2-place spread after pick 150 therefore isn't purely some sudden collapse of human football knowledge.

Although, having participated in late fantasy drafts, I would not rule that out entirely.

Your draft board should change jobs

I think this leads to a more useful way of thinking about rankings.

Early in the draft, your board should mostly help you price players.

The market is tight. Deviating aggressively requires a real reason.

Later in the draft, your board should help you find uncertainty.

Who is the market fighting about?

Why?

Where do you actually have a view?

That means I would approach preparation differently by section.

Inside the top 50, understand the major players, your roster construction and where your model meaningfully differs from the market.

From 50–100, start paying more attention to source disagreement and tier structure.

After 100?

That is where I would do the homework.

Find the five or ten players with the widest ranking ranges who you could realistically draft.

Then investigate the disagreement.

You don't need to form a strong opinion on every player.

Nobody needs a 207-player manifesto.

You need opinions in the places where the market does not already have one.

This might be where the actual edge lives

Fantasy managers love the idea that they will win a league because they correctly identified the seventh-best player instead of the ninth-best player.

Maybe.

But everybody knows the seventh- and ninth-best players are good.

The larger opportunity may come much later, where expert rankings are 20, 30 or 40 spots apart and the room is increasingly drafting from completely different boards.

That is where a small informational advantage can become a large price advantage.

And it explains one of the stranger parts of draft season.

We feel most confident early, when everyone agrees.

Then we maintain roughly the same confidence late, when almost nobody does.

Maybe we shouldn't.

The first two rounds are a highway with road signs.

By pick 100, we are driving through fog with six different GPS apps arguing about the exit.

You can still get where you're going.

Just stop pretending the map is equally precise.

Sources: FantasyPros — PPR Overall ADP; Auto Draft Years — 2026 consensus board. Board figures reflect the Auto Draft Years 2026 consensus board as generated on 2026-08-17. For informational and entertainment purposes only.