Why JSN, James Cook and the 49ers Make My Week 4 Card

Alright, here's what I like

First of all, yes, that's a cannoli in my pocket. You wanna talk about that or you wanna talk football? Because I've got three picks and San Francisco's injury report is already giving me agita.

I'm on Jaxon Smith-Njigba over 7.5 receptions, James Cook over 88.5 rushing yards, and the 49ers minus 2.5. One unit each. Three singles. We'll get to the 49ers, but let me enjoy the other two for a minute.

JSN just caught ten balls. Cook just ran for 154. Lovely. Everybody saw it, including the guy hanging the line. I want to know if Seattle keeps throwing to JSN and Buffalo keeps feeding Cook when Sunday's game stops cooperating. That's where these picks either make sense or make me look like a schmuck.

JSN is my favorite of the three. Cook comes next. San Francisco has the passing offense I want and a defensive injury list I absolutely do not. Let's get into it.

JSN: eight catches, not a highlight reel

The pick: Jaxon Smith-Njigba over 7.5 receptions at +105. Risk: 1.00u. Chargers at Seahawks, Sunday at 3:25 p.m. Central.

Seattle has opened the season by putting eight, nine and ten receptions on Smith-Njigba's game log. That's 27 catches and 405 receiving yards through three games. The attraction isn't just the yardage. It's that the offense keeps finding him.

Here's the detail I care about: Seattle's Next Gen Stats report credits him with 22 receptions while aligned out wide. Against Washington, eight catches on 12 targets came from wide alignments alone. So spare me the automatic slot-receiver explanation. The actual work is happening outside, too.

That gives me a cleaner reason to choose receptions rather than a touchdown or a yardage number. I'm backing repeated involvement. I don't need one 60-yard beauty that gets replayed all week. I need Seattle to complete eight passes to its best receiving argument.

He averaged seven catches a game in 2025. Seven catches? Beautiful game, Jaxon. My ticket still goes in the garbage. We need eight, and that's a lot to ask even when a guy is playing this well.

The obvious danger is Seattle getting ahead and taking its foot off the passing game. Another is a perfectly reasonable target count turning into seven completions. Coverage attention can send the ball elsewhere. Those are football reasons to lose, not bad-luck excuses I'll invent on Monday.

Derwin James and Donte Jackson were questionable on Friday. Neither is being treated as out in this pick. I like the role enough to make the call without pretending two defensive backs have already disappeared.

At +105, the price requires a 48.78% hit rate to break even. That's a property of the price, not my forecast. The case is continued involvement at the recorded terms. Three good Sundays do not entitle me to manufacture a probability.

Cook: feed the man

The pick: James Cook over 88.5 rushing yards at -115. Risk: 1.00u. Patriots at Bills, Sunday at noon Central.

Cook's first three games: 13 carries for 57 yards, 21 for 135, then 24 for 154. That averages 115.3. I can already hear somebody saying, 'Nori, the line's only 88.5, what the hell are we discussing?' We're discussing the 13-carry game, my friend. That one happened too.

I'm more interested in the carries. The last two games gave him 21 and 24 chances. Last season he averaged 18.2 attempts. That difference matters because I don't need him to keep ripping off six yards a carry if Buffalo keeps handing him the ball.

Twenty carries at 4.5 yards apiece gets us to 90. Eighteen carries at the same efficiency gets us 81 and a very annoying afternoon. Even 22 carries at four yards each leaves him on 88. The workload has to show up; the nice recent average cannot do the running for him.

New England ruled out Christian Barmore. Losing him helps the argument for workable rushing conditions. That's my football judgment, not a secret formula telling me his absence is worth an extra 11.7 yards.

Now the part that can make me look like an ass: Christian Gonzalez is out, too. Buffalo might decide the easiest way to move the ball is through the depleted secondary. A good matchup for the Bills is not a legally binding agreement to feed James Cook.

Ray Davis was questionable, not inactive. I have not quietly assigned his workload to Cook. Allen can take rushing opportunities, the Bills can fall behind, and ordinary blocking or tackling can make this harder than the pretty scenarios suggest.

I'm taking rushing yards because the opinion is about carries and ground production. A touchdown needs a different chain of events. Neither market is automatically better; this is the one that matches my argument at the saved price.

The -115 price needs 53.49% to break even. The bet rests on Buffalo giving Cook enough ordinary work, not on me pretending the last three games established a permanent 115-yard entitlement.

49ers: that half point is doing a job

The pick: San Francisco -2.5 at 1.85 decimal. Risk: 1.00u. Broncos at 49ers, Sunday at 3:25 p.m. Central.

Here we go. The pick that's going to have me standing six inches from the television asking a defensive coordinator why Denver is still on the field. I like San Francisco's passing offense enough to take -2.5. I wish the defensive injury report would stop getting in the way.

Purdy has completed 60 of 83 passes through three games for 789 yards, nine touchdowns and one interception. That's 72.3% completions and 9.5 yards per attempt. Useful evidence of an offense moving the ball, not a promise of three passing touchdowns every Sunday until Christmas.

The ATSKillers research snapshot also put San Francisco's quarterback, receiving and offensive-line units above baseline. That gave me several reasons to investigate the passing game rather than hitch the whole argument to one highlight. Denver's pass-defense research value was above baseline as well. There is resistance on the other side of the line.

Then comes the bill for the optimism. Nick Bosa, James Thompson Jr. and Mykel Williams were ruled out. Mike Evans and Dre Greenlaw were among the questionable players. Trent Williams had no game designation after practicing fully Friday.

If the depleted front lets Denver hang around, extend drives and keep scoring, San Francisco may need every bit of the offensive quality I'm backing. I haven't measured an exact point cost for those absences. Calling them a minor concern because I already like the favorite would be convenient bullshit.

I still prefer the 49ers at -2.5. At that number, a three-point win cashes. At -3 it pushes. At -3.5 it loses. Same teams, same final score, three different afternoons for your ticket.

If you show up with -3.5 and tell me we're on the same bet, we're gonna have a conversation. I took -2.5. A field-goal win pays this ticket. It buries yours. That's the whole point of caring about the number.

The recorded price is exactly 1.85 decimal: +0.85u profit on a one-unit win. About -118 is only an American-odds approximation. The exact decimal price governs this card, with a 54.05% price break-even rate.

One unit. The ego gets nothing.

JSN first, Cook second, San Francisco third. One unit risked on each, three units total. Preference order does not change the stake.

Could I call JSN a three-unit play and pound the desk? Sure. I could also wear two chains. Neither makes the football opinion better. I haven't established the win probabilities needed to justify confidence-based sizing, so it's one unit each.

These are singles in three separate games. That avoids a same-game parlay, but it doesn't make the thinking independent. Both props lean on recent usage. If I'm giving three games too much credit, the same mistake can bite twice.

A unit means risk, not targeted winnings. Keep any real-money decisions within a discretionary budget you can afford to lose. Nothing about a confident football opinion turns it into income.

Where the model fits

I use ATSKillers' team and player values to find the questions worth asking. Then I check the matchup, workload, injury information and recorded price. A high research value starts a conversation. It doesn't end one.

This card used the October 2 research snapshot. When the original selections were recorded Saturday morning, Sunday's official model decisions were still awaiting their evaluation window. These are my analyst picks, not advance announcements of what the model will select.

Those research values help compare the teams and players. They don't tell me the final score, and they aren't cover probabilities. I still have to look at who's playing, how the teams match up and what the sportsbook is asking me to pay.

Nori's results have their own record. Model wins don't pad it, model losses don't explain it away, and older Nori results aren't imported into this new NFL desk. You'll see the original call and what happened to it.

What gets checked after Sunday

The scoreboard gets first say. After that, I want to know whether the football argument held up. A winning ticket can come from lousy reasoning. A losing ticket can expose a bad assumption or a good opportunity that didn't finish. Neither gets a free pass.

For JSN, we check the catches, verified targets and participation, then ask how the game state affected the passing plan. If eight receptions arrive only after a frantic late chase, that is different from being involved all afternoon. Alignment evidence belongs in that review only if we can actually verify it.

For Cook, start with carries and the share of rushing work. Then look at efficiency and explosive runs. Did Buffalo give him the opportunities I expected, or did one enormous run bail out a thin workload? The total grades the pick; the distribution helps explain it.

For San Francisco, separate sustained offensive possessions from short fields, returns and unusual turnovers. A defensive touchdown can help cover -2.5 without proving a damn thing about my passing-game argument.

The same questions apply if the card loses. We don't move the goalposts because the result bruised my ego. One week cannot establish a profitable method, but it can show us exactly which assumptions deserve another look.

The terms stay on the ticket

The original card was recorded October 3 at 7:18:17 a.m. Central. Website publication is a separate timestamp shown at the top of this page. The BetMGM public quotes were observed that morning; their upstream refresh times were not exposed. They are recorded references, not executable offers or evidence that a wager was placed.

Weather and final Sunday inactives were not fully assessed when the card was frozen. Those limitations stay attached to it. Later news belongs in a dated follow-up, not a quiet rewrite that makes Saturday's decision look smarter.

Results will be recorded after checking final statistics and participation. Under this desk's stated convention, props require at least one snap; verified non-participation is void. Once there is action, an early exit does not earn an invented refund. Canceled or interrupted games stay unresolved until the disposition and applicable rules are established. This tracking convention is not a promise about every sportsbook jurisdiction.

The prices stay, the stakes stay, and the losses stay if that's how it goes. You can find me here after the games. If I go 0-for-3, I expect to hear about it. Just wipe the cannoli cream off your keyboard first.

Market references & reading

• Recorded JSN public quote
• Recorded Cook public quote
• Recorded Denver at San Francisco market
• Published BetMGM house rules used for desk convention

Follow the card