Team Intelligence · NFC West · 2026 preseason

Arizona Might Be the Most Reasonable Offense in Fantasy Football

The model forecasts a middle-of-the-league offense — 22.8 points a game, 18th — with its best fantasy real estate at running back (11th), and the market has it priced about right — the two disagree by 6 places, inside the bar this series sets before calling anything mispriced. Comparable offenses have supported 2.7 fantasy starters.

ADY
Auto Draft Years
17 August 2026 · 9 min read
22.8
proj points/game (18th)
2.7
expected fantasy starters
RB 11th
relative strength
+6
market vs model, in places

The short version

  • Arizona went 3–14 in an injury-ravaged 2025 and scored 20.9 points per game.
  • AutoDraftYears projects a modest rebound to 22.8 PPG, 18th in the NFL.
  • The bigger projected change is philosophical: pass rate falls from 64.0% to 57.7%.
  • RB is Arizona's best positional environment at 11th.
  • Comparable offenses produced 2.7 fantasy starters. Think three useful players, not six.
  • The market ranks Arizona's fantasy roster 12th; the model ranks it 10th. That is basically agreement.
  • McBride is appropriately expensive.
  • Do not reach for Love, especially after the ankle injury.
  • Watch Allgeier's price. He is the player most capable of becoming a buy if the market moves.
  • Sometimes the edge is noticing the market has already done the obvious homework.

Fantasy football has an uncomfortable relationship with the phrase do nothing.

We want sleepers. Fades. Breakouts. Someone ranked 78th whom the computer believes should be 31st, preferably so we can screenshot it in December and quietly delete the other 47 recommendations.

Arizona offers a less emotionally satisfying possibility.

The Cardinals were bad in 2025. Genuinely bad. 3–14, 21st in scoring, quarterback hurt, running backs disappearing into the medical tent like it had a loyalty program.

And now?

The AutoDraftYears model expects them to improve.

The fantasy market does too.

Unfortunately, everyone appears to have noticed.

That makes Arizona's real 2026 question less exciting, but probably more useful:

How much should we pay for the Cardinals' rebound when the rebound is already priced in?

2025 Wasn't Just Bad. It Fell Apart.

Arizona entered 2025 expecting to take a step forward.

Instead, it started 2–0 and then won one of its final 15 games.

The Cardinals finished 3–14 and scored 20.9 points per game. Jonathan Gannon was fired after the season.

But the final scoring number almost undersells how strange the season became.

James Conner suffered a season-ending foot injury in Week 3. Trey Benson, who was supposed to inherit the backfield, landed on injured reserve shortly afterward. By December, Arizona was discussing the season as essentially a lost year at running back.

Then Kyler Murray suffered a foot injury after five games and never returned. He finished with 962 passing yards and six touchdowns. The Cardinals eventually released him this March, ending his seven-year run in Arizona.

Marvin Harrison Jr. wasn't spared either. A concussion, appendicitis and injuries to both feet limited him to 12 games and 41 catches.

This is generally not the recommended recipe for offensive continuity.

And yet something weird happened.

Jacoby Brissett came in and the passing game became surprisingly functional.

Brissett threw for 3,366 yards and 23 touchdowns in 12 starts. Arizona finished seventh in the NFL in net passing yards per game despite ranking near the bottom in scoring.

Trey McBride went completely feral.

He caught 126 passes for 1,239 yards and 11 touchdowns, breaking the NFL record for receptions by a tight end. Michael Wilson quietly reached 1,006 receiving yards.

So Arizona's 2025 offense was not simply useless.

It was lopsided.

The running game collapsed. The original quarterback disappeared. Harrison couldn't stay healthy. Brissett spent much of the season throwing because Arizona was constantly chasing games.

Which explains an important number from the AutoDraftYears data:

Arizona threw on 64.0% of its snaps.

That wasn't necessarily an offensive identity.

Sometimes when your house is flooding, “use more towels” begins to look suspiciously like a strategic philosophy.

Arizona Cardinals: 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.

The Model Thinks Arizona Becomes…Normal

This is where 2026 gets less dramatic.

AutoDraftYears projects the Cardinals for 22.8 points per game, up from 20.9.

That moves Arizona from 21st in the model's 2025 scoring measure to an 18th-ranked projected scoring environment.

Not a revolution.

More like the franchise has successfully completed the paperwork required to rejoin the middle class.

Metric20252026 Projection
Points/Game20.922.8
Pass Rate64.0%57.7%
AutoDraftYears Volume59.7 snaps/g61.2
Red-Zone Trips3.6/g3.2/g
Overall Scoring Rank21st18th
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.

The bigger change is how the model expects Arizona to score.

Pass rate falls 6.3 percentage points.

That matters because Arizona spent the offseason rebuilding the exact part of the offense that vanished last year.

The Cardinals signed Tyler Allgeier, brought James Conner back and then used the third overall pick on Notre Dame running back Jeremiyah Love — the highest-drafted NFL running back since Saquon Barkley in 2018.

They also hired Mike LaFleur, formerly the Rams' offensive coordinator, as head coach. LaFleur will call the plays.

The model's positional environments reflect the change:

  • RB: 11th
  • QB: 21st
  • TE: 21st
  • WR: 22nd
Position environment
Arizona Cardinals 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.

Running back is clearly the best fantasy real estate on the roster.

Which would be extremely useful if the Cardinals had kindly provided us one obvious running back to draft.

They have not.

Offensive pie against opportunity concentration
Arizona — mid-sized 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 piedistributedARI
79% 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.

This Looks Like a Three-Starter Offense

The AutoDraftYears comps make the team-level expectation wonderfully boring.

Take Arizona's projected profile — 22.8 points, 57.7% pass rate, 61.2 snaps and 3.2 red-zone trips per game — and compare it with the 30 closest team-seasons from 2016 through 2025.

Those offenses averaged 2.7 fantasy starters.

Bear case: 2.

Base case: 3.

Bull case: 3.

2.7
expected fantasy starters
Bear2
Base3
Bull3

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 is one of the least ambitious bull cases ever published.

Across all 320 historical team-seasons in the AutoDraftYears sample, the average offense produced 2.6 fantasy starters. Arizona's comps produced 2.7.

So the model isn't forecasting a fantasy gold mine.

It sees roughly three players worth caring about.

This becomes important when we look at the prices.

Unfortunately, the Market Has Also Met Arizona

The market ranks Arizona's fantasy roster 12th.

The model ranks it 10th.

That's only a two-place gap, well inside the eight-place threshold AutoDraftYears uses before calling an entire offense meaningfully mispriced.

In other words:

The Cardinals are improving.

Fantasy managers know.

Please cancel the parade.

PlayerADPMarketModelAdjusted EdgeCall
Trey McBride24TE2TE20Fair Price
Jeremiyah Love26RB13RB16-6Fair Price
Marvin Harrison Jr.81WR34WR33+4Fair Price
Michael Wilson92WR42WR440Fair Price
Tyler Allgeier150RB50RB30+7Fair Price
Jacoby Brissett177QB28QB33—Thin Market
PlayerADPMarketModelEdgePPGRoleRiskCall
Trey McBrideTE25TE2TE201683MediumFAIR PRICE
Jeremiyah LoveRB28RB13RB19-9—81MediumFAIR PRICE
Michael WilsonWR92WR38WR43-61170MediumFAIR PRICE
Tyler AllgeierRB140RB47RB31+7750LowFAIR PRICE
Jacoby BrissettQB193QB27QB33—1357MediumTHIN MARKET
James ConnerRB212RB66RB84—554LowTHIN 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.

No buys.

No fades.

This is not the model refusing to have an opinion.

This is the opinion.

McBride Is Expensive Because He Is Trey McBride

McBride might be the cleanest example of the difference between liking a player and finding value.

ADP: 24.

Market: TE2.

Model: TE2.

Role Score: 83.

After 126 catches, everyone has discovered Trey McBride.

Congratulations to all involved.

If you draft him, you are paying an elite-tight-end price for someone the model also considers an elite tight end. That's perfectly reasonable.

But “good player” and “good price” are not synonyms.

I enjoy guacamole tremendously.

Chipotle has still managed to monetize this information.

Love Is the Player I Would Be Most Careful With

Love is harder.

Arizona didn't take a running back third overall because it imagines him operating a charming little timeshare.

The model gives him an 81 Role Score, ranks him RB16 and likes Arizona's RB environment considerably more than its passing environments.

But the market already has him RB13 around pick 26.

And now there is another complication.

Love suffered a high ankle sprain in Arizona's preseason game against Las Vegas and is expected to miss the rest of the preseason. The Cardinals remain hopeful he can return for the regular-season opener.

That doesn't make him an automatic fade.

It does make reaching even harder to justify.

Arizona currently has Allgeier, Love, Conner and Trey Benson on the depth chart. The franchise has described Love as the eventual top back, but Allgeier was signed before the draft with starting responsibilities in mind, and Conner is back from last year's foot injury.

Love may eventually dominate this backfield.

But pick 26 is expensive real estate for eventually.

Allgeier Is the Name I Keep Coming Back To

Then there is Tyler Allgeier.

Market: RB50.

Model: RB30.

At first glance, that looks like the exact sleeper headline we've been waiting for.

Sound the alarms. Wake the neighbors. Inform local government.

Except the model adjusts for the fact that rankings naturally get messier this deep in drafts.

After that adjustment, Allgeier's edge is only +7 positional spots.

The RB buy threshold is +13.

So he isn't a buy.

Yet.

And I actually like that answer.

His price is low enough that the situation can change faster than the cost does.

Love's ankle injury matters. Conner is returning from a significant foot injury. Arizona clearly wants to run substantially more than it did last season.

If Allgeier's ADP falls while his role remains intact, he is the Cardinal most likely to turn into an actual value.

Not because RB50 versus RB30 looks exciting on a graphic.

Because the price could eventually become cheap enough to compensate for the uncertainty.

NO BUY
Nobody clears the bar
The closest is Tyler Allgeier at +7 position places, short of the threshold. Forcing a buy here would be the template writing, not the model.
FAIR PRICE
Trey McBride
Trey McBride is priced at pick 25 (TE2); the model has him TE2. What fired: the model and the market land within a few position places of each other and a projected role scoring 83 of 100.
NO FADE
Nothing here is overpriced
No player on this roster is priced far enough ahead of the model to call a fade.

What Should Fantasy Managers Actually Do?

Mostly?

Stop trying to discover a secret Arizona doesn't contain.

McBride is an elite tight end at an elite-tight-end price.

Love is talented, highly drafted and sitting in the model's best Arizona positional environment. He is also already expensive, injured and surrounded by competent alternatives. Don't turn “I want him” into permission to reach.

Harrison and Wilson are essentially where the model expects them to be.

Allgeier is worth watching because his cost is low enough for new information to create an edge.

And that is probably the broader lesson here.

Arizona was a disaster in 2025, but it wasn't a talentless disaster. It lost its starting quarterback, its first two running backs and chunks of Marvin Harrison Jr.'s season, then rebuilt the coaching staff and backfield.

Of course improvement is plausible.

The problem is fantasy football does not award points for correctly identifying that a football team should improve.

You still have to beat the price.

Right now, Arizona's rebound is sitting on the shelf with a barcode already attached.

I know.

Disgusting.

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.