Team Intelligence · AFC East · 2026 preseason

The Jets Are Cheap. For Once, That's the Beginning of the Conversation.

The model forecasts a bottom-10 scoring offense — 17.8 points a game, 32nd — with its best fantasy real estate at quarterback (31st), and the market is still buying it as the 31st offense to own, 12 places behind where the model has it. Comparable offenses have supported 1.3 fantasy starters.

ADY
Auto Draft Years
17 August 2026 · 9 min read
17.8
proj points/game (32nd)
1.3
expected fantasy starters
QB 31st
relative strength
+12
market vs model, in places

The short version

  • 2025 was legitimately ugly: 3–14 and 29th in scoring despite starting the same offensive line all season.
  • The Jets cycled through Justin Fields, Tyrod Taylor and Brady Cook, while Garrett Wilson played only seven games.
  • 2026 is different: Geno Smith is QB1 and Frank Reich takes over the offense.
  • AutoDraftYears still projects only 17.8 PPG, basically unchanged from 17.6.
  • The important model change is volume: 59.5% pass rate, 58.5 snaps and 2.6 red-zone trips per game.
  • Historical comps average only 1.3 fantasy starters.
  • The Jets are 26th by the model and 22nd by the market — no team-level discount — and nobody clears the buy threshold.
  • Garrett Wilson is roughly fair.
  • Breece Hall is the fade: RB16 market versus RB38 model.
  • Watch red-zone trips early. If that number actually rises, the model may be underselling the offensive change.
  • The larger lesson remains wonderfully annoying: an undervalued offense does not automatically contain an undervalued player.

The New York Jets started the same five offensive linemen in every game last season.

This feels important because when an offense is terrible, offensive-line chaos is one of our favorite explanations. It is football's version of blaming traffic. Plausible. Difficult to disprove. Everyone nods.

Not here.

The Jets had their most stable offensive line since 2012. They still went 3–14, scored 17.6 points per game and finished 29th in scoring.

So before we discuss whether Garrett Wilson is a value or whether Breece Hall is overpriced, we should probably deal with the more basic question.

What exactly happened to the Jets?

2025: They found several different ways not to score

Aaron Glenn's first season started 0–7.

Justin Fields opened as the quarterback. He was hurt early, returned, struggled, got replaced by Tyrod Taylor at halftime of the seventh loss and eventually lost the starting job. Taylor later dealt with his own injuries. Brady Cook ended up throwing 153 passes.

By season's end, Fields, Taylor and Cook had combined to produce just 2,777 passing yards. Fields led the team with 1,259.

That is not really a passing offense.

It is three quarterbacks sharing custody of one.

The Jets finished with only 2,784 net passing yards and 15 passing touchdowns while taking 60 sacks. The weird bright spot was quarterback rushing: Fields, Taylor and Cook combined for 575 rushing yards, the most by Jets quarterbacks in a season in franchise history.

Useful in fantasy.

Less useful when your actual job is occasionally throwing the football.

And then Garrett Wilson got hurt.

Wilson had 36 catches, 395 yards and four touchdowns in seven games, then a knee injury effectively erased the rest of his season. He was placed on injured reserve in November and never returned.

This matters because before the injury, Wilson was still doing Garrett Wilson things inside a passing game that appeared to have been designed during a power outage.

Breece Hall, meanwhile, quietly produced 1,065 rushing yards at 4.4 yards per carry and another 350 through the air.

So 2025 wasn't a total absence of fantasy talent.

It was talented players trying to survive an offense that finished with 4,481 yards, 30 touchdowns and a negative-19 turnover differential.

New York Jets: scoring, pass rate and volume, 2021–2026
Solid line is points per game against the left axis; dashed lines are pass rate (right axis) and snaps from scrimmage.
The 2026 point on each line is the model's forecast, not an actual.

And this is where 2026 gets interesting.

Because the Jets did not simply run the same thing back.

They hired Frank Reich as offensive coordinator and traded for Geno Smith, who is now QB1.

New quarterback.

New play caller.

Healthy Garrett Wilson.

Same Breece Hall.

Surely the model loves the bounce-back.

Well.

About that.

What the model thinks happens next

AutoDraftYears projects the Jets to score 17.8 points per game in 2026.

Thirty-second in the NFL.

Last year?

17.6.

So after changing quarterbacks and offensive coordinators, the model has generously awarded New York two-tenths of a point per game.

Please spend it responsibly.

The broader offensive profile does change, though.

Metric20252026 Model
Points per game17.617.8
Pass rate52.1%59.5%
Snaps per game55.958.5
Red-zone trips/game2.22.6
2025 to 2026
Where the forecast moves
Percentage change against last season's actuals.
Bars are the forecast's own movement; the roster changes above are what the model is responding to.

That pass-rate jump matters.

The 2025 offense leaned heavily on the running game, including its quarterbacks. The 2026 model expects Reich and Smith to produce a much more conventional passing environment.

More plays.

A lot more passes.

More red-zone opportunities.

Just not substantially more points.

That's an unusual combination, but it also explains the model's strange relationship with this roster.

The Jets' best projected positional fantasy environment is quarterback.

It ranks 31st.

Running back ranks 32nd. Wide receiver 31st. Tight end 31st.

Sometimes being the strongest part of something is not the compliment it appears to be.

Someone has the nicest food court restaurant at the airport.

Position environment
New York Jets 2026 grades by position
Bars are inverted league rank, so taller is better. Green is top-10, red is bottom-10.
An environment grade is a thumb on the scale for a close call, not a ranking of the players in it.
Offensive pie against opportunity concentration
New York Jets — small pie, concentrated
Horizontal: how many fantasy points this offense has to give out. Vertical: how tightly its projected points sit on its top two skill players.
Small pieconcentratedLarge pieconcentratedSmall piedistributedLarge piedistributedNYJ
100% of this offense's projected skill points sit on its top two players, against a league median of 63%. Worth being precise about what this axis does: across 320 historical team-seasons concentration has almost no relationship with how many fantasy starters an offense produces (r = −0.07). The size of the pie decides how many; concentration only decides who.

History isn't especially excited either

AutoDraftYears matched the 2026 Jets profile against the 30 most similar offenses from 2016–2025.

Those teams averaged only 1.3 fantasy starters.

The distribution was basically:

Bear case: one. Base case: one. Bull case: two.

Comparable offenses produced:

  • a top-12 QB only 3% of the time;
  • two top-24 RBs 43% of the time;
  • three top-36 WRs 60% of the time;
  • a top-12 TE only 13% of the time.
1.3
expected fantasy starters
Bear1
Base1
Bull2

Read off the 30 most similar offenses in 2016–2025, not asserted. A starter is a QB1, an RB2, a WR3 or a TE1 — top 12, 24, 36 and 12 at the position. Across all 320 team-seasons the average offense produced 2.6.

That does not mean Geno Smith cannot make Garrett Wilson better.

He probably can.

It does not mean Breece Hall cannot have a useful fantasy season.

He probably will.

It means there is a difference between improving from the 2025 Jets and becoming a good fantasy ecosystem.

That is a very low bar.

You can climb out of the basement without reaching the penthouse.

So is there a Jets discount hiding anywhere?

Here's the strange part.

When every NFL roster is converted from player rankings into expected fantasy value, the market ranks New York 22nd.

AutoDraftYears ranks them 26th.

Four places apart, with the market slightly the more generous of the two.

The threshold for calling a whole roster mispriced is eight.

So, no.

Buy Jets?

Nope.

Not one player currently clears the model's buy threshold.

This is my favorite kind of model result because it refuses to give us the convenient answer.

There is no bulk discount at the door, and there is nothing on clearance inside either.

Costco can price the entire warehouse correctly.

You still don't need seven televisions.

PlayerADPMarketModelEdgePPGRoleRiskCall
Breece HallRB32RB16RB39-26977MediumFADE
Garrett WilsonWR39WR18WR18+31277LowGOOD PLAYER, BAD ROOM
Kenyon SadiqTE180TE22TE29——75MediumTHIN MARKET
Braelon AllenRB186RB58RB134—339LowTHIN MARKET
Adonai MitchellWR190WR72WR153—554MediumTHIN MARKET
Mason TaylorTE273TE38TE47—657LowTHIN MARKET

Market and Model are ranks within the position. Edge is the model's position rank against the market's, after removing the drift that affects every player at that price — so it reads as “the model likes him this many places more than it likes the typical player who costs this much.” A call needs the edge to clear the position's bar (QB 5, RB 15, WR 11, TE 5 places), set at two-thirds of that position's own spread.

Model against market
Disagreement, in position places
Bars are the de-drifted edge — the same number the calls in the table are made from.

Garrett Wilson: good player, questionable neighborhood

Wilson is the player I am most willing to draft.

His ADP is 39, or WR19.

The model has him WR18.

After adjusting for normal price-level model drift, his edge is only +4 positional spots.

The WR buy threshold is 11.

So Wilson is not a model buy.

He is also not a fade.

He has a 77 Role score, Low risk and a roughly 12 PPG projection.

And unlike last August, we have a fairly obvious reason to believe his situation could improve: he is healthy again, Geno Smith has replaced the Fields/Taylor/Cook rotation, and Frank Reich is running the offense.

I like Wilson.

I just don't need to turn liking Wilson into a larger philosophical commitment to the New York Jets.

WR19 is fine.

If he falls, I become interested.

Breece Hall is different

Hall's ADP is 34.

Market rank: RB16.

Model rank: RB38.

Adjusted edge: -24 positional spots.

That is a fade.

Hall's 77 Role score is the obvious counterargument, and his 2025 season showed that he can still generate production inside a bad offense. He ran for 1,065 yards despite everything happening around him.

But that's also part of the problem.

We already saw a pretty healthy workload.

We already saw good rushing efficiency.

And the result still came in an offense that struggled badly to create touchdowns.

Now you're being asked to pay RB16 prices while AutoDraftYears gives New York the 32nd-ranked RB environment and projects Hall around 9 PPG.

Fantasy managers have spent several seasons waiting for the complete Breece Hall Experience.

I understand.

The talent makes it easy to imagine.

But “maybe everything finally comes together” is also one of fantasy football's most successful subscription businesses.

At pick 34, somebody else can renew.

NO BUY
Nobody clears the bar
The closest is Garrett Wilson at +3 position places, short of the threshold. Forcing a buy here would be the template writing, not the model.
FADE
Breece Hall
Breece Hall is priced at pick 32 (RB16); the model has him RB39. What fired: the market pays for 26 position places more than the model gives him and a 32nd-ranked RB environment behind him. Cutting against it: a projected role scoring 77 of 100.

Everyone else belongs in the garage-sale section

Omar Cooper is around WR66.

Kenyon Sadiq, TE22.

Adonai Mitchell, WR74.

Geno Smith himself is QB30 at pick 186.

I don't have especially strong objections to any of them.

I also don't have especially strong opinions.

At these prices we are no longer conducting precision asset valuation. We are standing in someone's driveway holding a lamp and asking why it costs $6.

Take your swings.

With no discount to be had at the top of this roster, these cheap bets are the only place left where the price is doing any work.

Just don't pretend WR74 versus WR97 is a distinction humanity has successfully mastered.

What I would watch in September

There are two numbers that could make this projection look too pessimistic.

The first is pass rate.

AutoDraftYears expects New York to jump from 52.1% to 59.5%. If Reich actually turns this into a significantly more aggressive passing offense, Wilson's environment changes quickly.

The second is red-zone trips.

The Jets managed only 2.2 per game in 2025.

The model projects 2.6.

That's the number I care about most.

If New York is regularly approaching three red-zone trips per game, then perhaps Geno Smith and Reich have done more than rearrange the furniture.

If they're still living around 2.2?

We probably know this movie.

What should fantasy managers actually do?

Don't draft the Jets looking for a team-level bargain. The market has them 22nd and the model 26th, which is the two of them agreeing.

That is useful information.

It just isn't an invitation.

The 2025 Jets were genuinely bad, but they were also unusually dysfunctional at quarterback and lost Garrett Wilson after seven games. The 2026 version has legitimate reasons to be different.

The model simply does not believe different automatically means good.

That leaves me comfortable with Wilson at the right price.

Interested in cheap late-round swings.

And unwilling to pay RB16 prices for Hall.

The Jets may be an undervalued football roster.

They may even become a better offense.

But fantasy drafts don't let us buy the New York Jets ETF.

We have to buy the individual stocks.

And right now, most of them are still priced exactly where I don't want them.

Was this useful?

Sources: Auto Draft Years FF Model workbook — Team Data By Year, Team Stat Ranking, 2026 Team Context; Auto Draft Years 2026 consensus board (16 authoritative market sources); Player outcomes 2016–2025, Pro Football Reference via the project dataset. Team environment grades, forecasts and player model ranks are proprietary. Every figure on this page is generated from the dataset at build time; none is typed in by hand.