One week of actual football is in, waivers have churned, and the Power Rankings no longer have to run on vibes and offseason rank math alone. There are results now. Real ones, with decimal points and everything.

Standings are split clean down the middle. Five teams at 1-0 (Austin Clout Demons at 170.76 points for, What r u doing step-substation? at 120.46, JahJonJuwan Jameson at 113.24, Tanjiro Kamado at 110.56, My Njigba Hurts at 97.62) and five at 0-1.

The seeding is already being weird about it. San Diego Football Team is 0-1 and sits at the 6 seed, ahead of Skatteboys, Tony's Titans, Rat Kings and The Disturbance Regime, because they dropped 142.20 points and ESPN breaks it on total scoring. A team that lost is outranking all four other teams that also lost. Best loser in The Dirty P. It's a title. Nobody wants it.

Hold that thought. The rankings have an opinion about San Diego, and it's a new one, because the rankings are new.

---

How the Power Score actually works, and why it works differently than it did last week (click to expand)

I rebuilt the model. All of it. Before anybody accuses me of moving the goalposts because my own team fell, let me get ahead of that by showing you exactly what was broken.

The old Power Score had two real problems.

Problem one: it thought quarterbacks were roughly three times more valuable than they are. The old system converted every player's consensus rank into a score on one shared curve. Trouble is, QB rank and RB/WR/TE rank come off boards of wildly different depth. QB12 in a 10-team league is a guy somebody probably dropped. WR12 is a first-round pick. Running both through the same curve quietly handed every roster with a name-brand quarterback a bonus it hadn't earned, and shortchanged everybody who spent early capital at the positions that actually get started twice a week.

Problem two: the Result Bonus graded you on your opponent's performance. It ran on margin of victory. Margin is two teams' scores subtracted from each other, which means half of your grade was somebody else's Sunday. San Diego scored the second-most points in the league in week 1 and the old model docked them 21.4 for it, because the one team that scored more happened to be sitting across from them. That's not a measurement. That's a coin flip with a spreadsheet attached.

So here's the replacement, which runs on points instead of rank positions.

Step one: every player gets a real point projection

Not a rank. An actual rest-of-season point total, blended 50/50 from two independent sources. ESPN's own rest-of-season projection, which is already computed in this league's exact scoring settings, and FantasyPros' rest-of-season consensus, mapped onto the same points ladder position-neutrally so the conversion doesn't secretly favor one position over another. Quarterbacks use FantasyPros' published rest-of-season point projection directly.

The two sources agree well enough to be worth blending rather than picking a favorite (0.87 correlation on QB points, 0.89 on RB/WR/TE order). When two boards built by different people using different methods land in the same place, that's the closest thing to confirmation this hobby offers.

Step two: optimal lineup, same eight slots as before

One QB, two RB, two WR, one TE, two FLEX, built off your current, post-waiver roster. Not the lineup you actually set on Sunday. The best one available to you, because this measures what you own rather than what you clicked.

The fix here is quiet but total. A great quarterback and a great running back are now compared in the same units, projected points, instead of being two separate ordinal boards run through one formula and hoping for the best. Josh Allen still comes out enormous, 365.7 rest-of-season points, the single largest value on any roster in The Dirty P. He just comes out enormous for a reason you can check rather than because the curve was tilted.

Step three: a depth bonus that's measured, not guessed

Your best three bench players earn you 20% of whatever they're worth above replacement level. Replacement level isn't a number I picked. It's pulled straight from what this league actually did on waivers this week, the real points-per-position of the players who just got added (QB 279.3, RB 77.0, WR 124.8, TE 116.7, rest-of-season pace). The 20% reflects that byes and injuries force roughly 1.6 bench players into your lineup per week across a full season. Your bench matters. It does not matter as much as your starters.

Step four: results get a vote sized to what one game is worth

Actual points scored. Not margin. Never margin again.

How much that vote counts got derived from this league's own week 1 scoring variance rather than chosen by feel, and it comes to about 8% of the final number after one game. That climbs every week as results pile up. By November it will be doing most of the talking.

The worked example

San Diego Football Team, start to finish:

• Optimal eight: Brock Purdy 299.4, Bijan Robinson 295.6, David Montgomery 192.1, Amon-Ra St. Brown 252.6, Zay Flowers 201.6, Harold Fannin Jr. 132.3, Garrett Wilson in the flex at 193.9, TreVeyon Henderson in the other flex at 147.2. That's 1714.7 rest-of-season points.
• Divide by 16 games: 107.2 per game.
• Depth bonus: 28.7 rest-of-season points, third-largest in the league, worth about 1.8 a game. Running total 109.0.
• Week 1: they scored 142.20. At roughly 8% weight, that's 109.0 + 0.08 × (142.20 − 109.0), and you land on 111.7.

That's the Power Score. It's expressed in points per game, so it's directly comparable to a real weekly score instead of living on some abstract four-digit scale. Everything on this board sits between 102.2 and 114.7, which is a narrower spread than the old system produced, and that's the point. This league is closer than last week's numbers made it look.

Where it's still soft

Two places, and I'd rather name them than have somebody find them.

My Njigba Hurts is genuinely uncertain, not just ranked low. A.J. Brown is on ESPN's injury report as IR-eligible. ESPN's point projection reflects that. FantasyPros' rank doesn't fully. Run this same model on FantasyPros alone and that team ranks 7th instead of 9th. Two spots of swing sitting on one player's medical status is more exposure than any other roster in the league carries right now.

The results weight is a judgment call dressed as a derivation. 8% came out of this league's own variance, which is one week of data from ten teams. It's a defensible number. It is not a precise one, and anyone who tells you they can extract a stable variance estimate from ten data points is selling something.

Now the countdown.

---

10. The Disturbance Regime, 102.2

Dead last preseason, dead last now, and the diagnosis is not what you'd expect from a team in this position. The starting eight is fine. James Cook III at 260.6, Chase Brown at 246.4 and Justin Jefferson at 237.9 is a real foundation, and it adds up to 104.3 a game, ahead of Skatteboys, Rat Kings and My Njigba Hurts on starter value alone.

The bench is where it falls apart. Depth bonus of 3.6, and the next-thinnest team in the league is at 17.8. That's not a gap. That's a different category of roster construction. One bye week, one hamstring, and this lineup starts leaking points it has no way to replace.

They were the busiest team in the league on waivers, which tracks. Out: Isaac TeSlaa, Kaleb Johnson, and Rashid Shaheed. In: Chris Brooks (RB52 week 2), Khalil Shakir (WR49 week 2, WR54 rest-of-season), and the one that matters, Jordan Love (QB16 week 2, QB17 rest-of-season).

Love is there because Kyler Murray started week 1 and finished at -0.38, then cratered to QB33 for week 2 even as his rest-of-season rank held at QB21. Here's the update: on real points, Love (276.5) is already the optimal-lineup quarterback on this roster. The board has made the call. Whether the manager agrees is a separate question with a lineup deadline attached.

9. My Njigba Hurts, 102.8

The biggest fall in the league, in either direction, and it's not close. Second in the preseason countdown, ninth now. Minus seven. They won their week 1 game. They're still ninth.

Last week's model blamed this on depth, and last week's model was wrong about that. Their depth bonus is 19.6, which is not great but sits comfortably ahead of Austin's 17.8 and in a different universe from The Disturbance Regime's 3.6. Depth isn't the problem.

The starting eight is. 102.0 per game, the lowest starter value in the league. Jaxon Smith-Njigba at 252.9 is the third-most valuable receiver on any roster in The Dirty P, behind only Ja'Marr Chase and Puka Nacua. Ashton Jeanty at 246.4 is a legitimately good running back who scored 29.7 on a 73.5% snap share in week 1. Then it stops. Luther Burden III at 166.1 and A.J. Brown at 155.9 are the WR2 and the flex, and Brown is carrying the injury-status question flagged above. Quinshon Judkins at 181.0 is the RB2.

Week 1 backed that shape up. Colston Loveland started and scored 0.0. Marvin Harrison Jr. put up 3.8 from the bench in a week where one national writeup declared the alpha-receiver dream over. They made no waiver moves.

Two stars and a long quiet stretch behind them. Last week I could tell this manager the drop was the model's fault. I can't do that this time.

8. Rat Kings, 103.1

Eighth preseason, eighth now, the second of the two teams that finished exactly where they started. Unlike the one below them, this roster is interesting.

Here's the thing: this is a deep roster. Depth bonus of 32.9, second-largest in the league behind only Tanjiro Kamado. The starting eight is the issue at 103.3 a game, ninth in the league, which is strange for a team rostering Joe Burrow (306.0), Puka Nacua (266.1) and Saquon Barkley (246.2). The top is fine. The middle is where Travis Etienne Jr. (193.0), DJ Moore (169.4) and Mike Evans (167.1) all land in the same narrow band, and a lineup full of similar guys doesn't separate you from anybody.

Then week 1 happened, and the 77.46 they posted is the second-biggest drag any team took in this model. Nacua, the consensus WR1 for week 2, managed 9.9 against a 17.38 projection. Barkley went 8.5 against 16.40. Two cornerstones, 18.4 points between them, while DJ Moore quietly put up 18.5 by himself from the bench. Their waiver response was Devaughn Vele (WR50 week 2, WR67 rest-of-season) in, Jayden Reed out. Not a needle-mover. A shuffle.

7. Skatteboys, 103.9

Third preseason, seventh now, down four, and I owe this team a correction.

Last week's model implied Skatteboys were being punished for having a bench and no stars. Under real points, that's backwards. Their depth bonus is 27.5, tied for fourth-largest in the league, ahead of JahJonJuwan Jameson's and step-substation's. This is a genuinely deep roster. It was deep last week too. The old math just couldn't see it.

The actual problem is the front. 103.4 a game from the starting eight, eighth in the league. Lamar Jackson at 331.6 is the second-most valuable quarterback in The Dirty P and Brock Bowers at 177.7 is the second-most valuable tight end, which is a fantastic pair of positional edges to own. Behind them, Bucky Irving (200.1) and Cam Skattebo (196.0) are the backfield, and neither one is a top-15 running back value in this league. CeeDee Lamb at 235.5 is real. Jaylen Waddle at 175.5 in the WR2 slot is fine and nothing more.

Worth sitting with for a second, because two teams that looked like the exact same story last week turn out to have opposite explanations. Skatteboys are deep with a soft middle in the lineup. My Njigba Hurts are top-loaded with a thin bench. Rank curves flattened them into one narrative. Points pulled them apart.

The 23.48-point loss to JahJonJuwan Jameson didn't move the needle much either direction. They started MarShawn Lloyd into a 3.7 (11.9 projected) and had Waddle post 0.7 from the bench, the rare bench call that was correct and still felt bad. On waivers they took Kenyon Sadiq (TE30 week 2, TE23 rest-of-season) and dropped Terrance Ferguson.

6. Tony's Titans, 105.8

From first in the preseason countdown to sixth. Down five. Nobody is going to accuse me of rigging this thing in my own favor.

The roster is fine. It's more than fine. De'Von Achane (236.0) and Breece Hall (214.0) in the backfield, Nico Collins (219.2) leading a four-deep receiver group with Drake London (197.0), George Pickens (194.7) and Christian Watson (184.6), Sam LaPorta (138.0) at tight end. Depth bonus of 27.5, tied for fourth in the league. It's a balanced, well-built team whose best non-quarterback, Achane, sits 117 points behind the league's best. This is a model that pays for top-end value, and there isn't any here.

The week 1 score is the rest of it. 85.16, third-lowest in the league, and the single largest reason this number reads 105.8 instead of something closer to 107.6.

One thing the new math settled. Jayden Daniels (311.2) beats Trevor Lawrence (290.8) on real projected points. The old rank board had Lawrence narrowly ahead at QB7 to Daniels' QB8, which produced a genuinely annoying subplot across two straight articles where the "optimal" lineup kept disagreeing with the obviously correct real-world call. Points settle it. Daniels was right. He's still right. That one's closed.

No waiver moves, which after last week reads as either patience or paralysis.

5. What r u doing step-substation?, 108.1

They won by 43.00, the biggest margin of week 1, and under the old model that blowout bought them the largest results bonus in the league. Under this one the margin is irrelevant and only the 120.46 they actually scored counts, which is a good number rather than a historic one. Net effect: they slide from 4th to 5th and nobody should lose sleep over it.

The positional edges are the story and they are excellent. Trey McBride at 191.1 is the most valuable tight end in the league by a comfortable 13 points over Brock Bowers, which in a one-TE format is a weekly head start over eight other managers. Caleb Williams (314.8) is QB4 for week 2 and rest-of-season after a 37.26-point opener that led every player in The Dirty P in overperformance. Christian McCaffrey at 267.9 is the fourth-most valuable running back in the league.

The waiver add already earned its spot. Jalen Coker (WR31 week 2, WR38 rest-of-season) comes in at 150.8 and slots directly into the optimal flex, with Isiah Pacheco going the other way. Adding a guy on Wednesday who immediately cracks your best eight is the whole point of the exercise.

Starter value of 105.5 puts them in the middle of the league, so this isn't a juggernaut. It's a team that's better at three specific positions than anybody else and average everywhere else, which in a ten-team league plays just fine.

4. Tanjiro Kamado, 109.7

A small piece of transparency before the analysis. An earlier pass of this rebuild briefly had Tanjiro first, the same place the old broken model had them. Fixing the quarterback weighting dropped them to fourth. I'd rather show you that than quietly ship the number that happened to agree with last week. Sixth preseason, fourth now, up two.

They're still very good. Ja'Marr Chase at 277.5 is the most valuable receiver in The Dirty P, and he scored 2.2 points in week 1 against a 16.4 projection while getting named in national recaps for two catches and twelve yards. Tanjiro Kamado won that game anyway, because Kenneth Walker III (255.0) went 32.6 and dragged the whole thing across the line. Drake Maye at 318.8 is the third-most valuable quarterback in The Dirty P. Chris Olave (208.3), Emeka Egbuka (192.2) and Parker Washington (176.4) make the receiver room four deep.

Two real weaknesses. Mark Andrews at 124.9 is the single lowest starter value on any optimal lineup in the entire league, which is a genuine weekly hole that a rank curve was more forgiving about than actual points are. And Jadarian Price at 167.2 is the RB2, which is thin for a contender.

The counterweight: the largest depth bonus in the league at 34.1. Whatever goes wrong here, there's an answer on the bench. No waiver moves, because there was nothing urgent to fix.

3. JahJonJuwan Jameson, 111.6

Ninth preseason to third, up six, tied with Austin for the biggest climb on the board. And here's the number that should worry everybody: 109.7 per game from the starting eight, the best in The Dirty P. Not second. First. Ahead of Austin, ahead of San Diego, ahead of everybody.

Jahmyr Gibbs at 353.2 is the most valuable non-quarterback in the league, more than 57 points clear of the next running back, and he went for 31.1 points on 34 touches in week 1 on his way to getting called a fantasy cheat code in national writeups. Jonathan Taylor at 284.9 is the third-most valuable running back in the league. That's a backfield nobody else can match, in a format that starts two of them and lets you flex a third.

So why third instead of first? Week 1. They scored 113.24, which is essentially exactly what this model projects them to score, so the results blend gave them a grand total of +0.15. San Diego's blend gave them +2.73. That 0.1 separating second from third is entirely one Sunday's box score, and it will be gone in a week.

The soft spots are real but survivable. Dalton Kincaid (129.7) at tight end is the second-lowest starter value in the league, and the flex slots lean on Jameson Williams (163.8) and Matthew Golden (148.5). Patrick Mahomes at 289.6 is the ninth-most valuable of the ten starting quarterbacks in this league, which is a sentence that would have gotten you laughed out of a draft room three years ago. Quiet waiver week: Emmett Johnson (RB53 week 2, RB52 rest-of-season) in, Jared Goff out.

2. San Diego Football Team, 111.7

The best loser is now the second-best team in the league, and this is the entry where the model rebuild pays for itself.

Under the old system, San Diego was seventh, punished 21.4 points for losing by 28.56 to the team that put up the highest score anybody has posted in The Dirty P this season. Under this one, their result is graded on the 142.20 they actually scored, second-highest in the league, and the blend hands them +2.73, the second-biggest results boost on the board. That single change is the difference between seventh and second. Same roster, same Sunday, honest arithmetic.

Fifth preseason, second now, up three. Bijan Robinson at 295.6 is the second-most valuable non-quarterback in the league. Amon-Ra St. Brown (252.6) is the fourth-most valuable receiver. David Montgomery is still the engine in his own backfield after a 27.4-point debut and a 23-to-10 touch edge, and Garrett Wilson (193.9) and Zay Flowers (201.6) fill out a receiver room that's three deep before you get to the flex.

They were also the most productive team on waivers in terms of actual lineup impact. Brock Purdy (QB7 week 2, QB10 rest-of-season) in, Bo Nix out. Nix had been on their bench posting 5.44 against a 16.1 projection, so no sentimentality was required. Purdy at 299.4 is immediately the starting quarterback. They also grabbed Travis Kelce (TE9 week 2, TE13 rest-of-season) and cut Jake Ferguson, and here's the wrinkle: on blended points the optimal lineup still wants Harold Fannin Jr. at 132.3 in the tight end slot over Kelce. That's a sentence I did not expect to write this year, and it's also why Kelce on the bench pushes their depth bonus to 28.7, third in the league. Good problem.

One real cloud. Flowers went for 150 yards and a touchdown on his way to 23.5 points and then exited with a hamstring. Multiple outside sources flagged his week 2 availability, and the consensus agrees: WR11 rest-of-season and WR32 for week 2, which is exactly what a healthy-long-term, questionable-right-now player looks like on a board.

1. Austin Clout Demons, 114.7

Seventh in the preseason countdown. First now. Up six, tied with JahJonJuwan Jameson for the biggest move in the league, and three full points clear of second place on a board where nobody else is separated by more than 2.3.

Two things got them here, and both are things the old model was actively bad at seeing.

Josh Allen is priced correctly for the first time. 365.7 rest-of-season points, the single largest value on any roster in The Dirty P, roughly 34 points clear of the next quarterback. The old system flattened him onto the same curve as every other startable QB and treated a genuine positional edge as a rounding error. He is not a rounding error. He put up 35.66 in week 1 as the consensus QB1 for both week 2 and rest-of-season, and this model finally says so out loud.

The running backs are real behind him. Derrick Henry at 253.6 with the full goal-line role intact on a 54.4% snap share. D'Andre Swift at 203.1 in the flex after going 31.9 against an 11.41 projection. Jeremiyah Love at 196.2 in the other flex. Javonte Williams at 214.0. That's four backs worth starting, in a format that lets you start four backs.

Then there's the 170.76, the best single-team week anybody has posted, with three starters clearing 30 points. At 8% weight it's worth +5.02 to their Power Score, comfortably the largest results boost in the league, and it's the margin between first and second.

Now the part their manager should actually worry about. Their depth bonus is 17.8, ninth out of ten. That looks impossible for a team that stranded roughly 50 unused points on its bench in week 1, until you notice that Isaiah Likely (141.3), the bench guy who scored 23.8 while the consensus held him at TE8, is now the starting tight end in the optimal lineup. Once you put the good bench player where he belongs, what's left behind him is close to nothing. And Henry's rest-of-season rank sits at RB9 against that RB4 week 2 number, which is the consensus quietly pricing in the calendar.

No waiver moves, which is defensible when you just scored 170. It will be less defensible in October.

---

Week 2 previews

What r u doing step-substation? (1-0) vs Tony's Titans (0-1)

ESPN's own current-lineup projections have this at 107.1 to 107.2. One tenth of a point. Not a typo.

Tony needs a bounce-back and drew the team that just won by 43. No soft landing anywhere in it. step-substation walks in with Caleb Williams and the league's most valuable tight end in Trey McBride, against a 6th-ranked roster carrying the third-lowest week 1 score in The Dirty P. Two teams separated by 2.3 points of Power Score and one tenth of a point of projection. This is the definition of a coin flip that both managers will describe as a must-win.

The lineup subplot has changed shape. Last week's rank board framed it as Watson-or-Pickens, with the consensus narrowly preferring Pickens. On real points, the optimal lineup starts both of them, George Pickens (194.7) and Christian Watson (184.6) in the two flex slots, with Drake London (197.0) alongside Nico Collins at receiver. The model isn't asking Tony to choose anymore. It's asking him to actually field the guys, which is the exact thing that cost him the margin last week.

The Disturbance Regime (0-1) vs Rat Kings (0-1)

The basement game, and now it's literally the bottom of the Power Rankings too, 10th against 8th. These two posted the league's two lowest scores in week 1 (76.52 and 77.46) and somebody has to come out at 1-1. Projection is 100.4 to 102.8, a polite way of saying neither side is expected to clear last week's 108.37 league average.

The subplot is the quarterback slot, and the model has already voted. Jordan Love (QB16 week 2) is now the optimal-lineup starter on The Disturbance Regime at 276.5 points, sitting next to Kyler Murray, whose week 2 rank cratered to QB33 after the -0.38 while his rest-of-season number held at QB21. Start Love and you're agreeing with the board. Start Murray and you're betting the market overreacted to one catastrophic afternoon. Either way the decision is visible from space.

Rat Kings added Vele, dropped Reed, and bring the second-deepest bench in the league into a game where depth is worth exactly nothing, because nobody's on bye yet. They'll have to win this one with Burrow, Nacua and Barkley showing up, which is not a thing that happened last Sunday.

San Diego Football Team (0-1) vs Tanjiro Kamado (1-0)

Best game on the board for signal, and the biggest numbers too. San Diego's 113.6 is the highest projected total of any team this week, and the 220.3 combined is the fattest matchup projection on the slate. They're favored 113.6 to 106.7.

For once the projection and the rankings agree with each other. San Diego is 2nd in the Power Score at 111.7, Tanjiro Kamado is 4th at 109.7, and ESPN has the 2 favored over the 4. Under last week's model, this was the 7th-ranked team being favored over the 1st, which was a flashing red light nobody wanted to look at. Fixing the results math fixed the contradiction.

San Diego's new pieces debut here. Purdy replaces the Bo Nix experiment, Kelce arrives as expensive bench insurance, and Montgomery is coming off 27.4. A roster that scored 142.20 and then got better. The Flowers hamstring is the one variable nobody controls, and it's sitting on this side of the field.

Tanjiro Kamado counters with Walker (255.0, coming off 32.6) and a Ja'Marr Chase who is either about to explode or about to post the most expensive two-catch afternoon in league history for a second straight week. Chase is still the most valuable receiver in The Dirty P by the model's own numbers. Mark Andrews at tight end is the soft slot and San Diego doesn't have a great answer at the position either, so that one's a wash.

Win here and San Diego is 1-1 with the best two-week scoring résumé in The Dirty P and a live argument for the top line. Lose and Tanjiro is 1-1 with the league's deepest bench and nothing else going right.

Skatteboys (0-1) vs My Njigba Hurts (1-0)

The two teams that fell hardest in this rebuild, playing each other. My Njigba Hurts dropped seven spots, Skatteboys four, and as covered above, for completely different reasons that last week's math lumped together.

The model and the standings disagree loudly here. My Njigba Hurts are 1-0 at the 5 seed. Skatteboys are 0-1 at the 7 seed. ESPN has Skatteboys favored 104.0 to 91.3, the widest projected gap of the week, and that 91.3 is the lowest projected total of any team on the slate. So the 1-0 team is projected to lose by nearly 13 to the 0-1 team, and the Power Score backs the projection, 103.9 to 102.8.

Jeanty and Smith-Njigba are the reason to believe in My Njigba Hurts, and they're a genuinely strong top two. Loveland's 0.0 and Harrison Jr.'s 3.8 are why the projection sits where it does, and the A.J. Brown status question is the swing factor that could make this ranking look wrong in either direction. Skatteboys added Sadiq at tight end, have Lamar Jackson and Brock Bowers providing weekly positional edges at two different spots, and need MarShawn Lloyd to beat 3.7 this time.

JahJonJuwan Jameson (1-0) vs Austin Clout Demons (1-0)

The marquee, and now it's the actual marquee. First against third. The two biggest risers in the league, both up six spots, both 1-0, playing each other in week 2. Austin's 111.6 is the second-highest projected total on the slate behind San Diego's, against JahJonJuwan Jameson's 106.5.

Gibbs (353.2) and Jonathan Taylor (284.9) against Josh Allen (365.7) and Derrick Henry (253.6). The most valuable non-quarterback in The Dirty P against the most valuable player in The Dirty P full stop. JahJonJuwan Jameson owns the best starting eight in the league at 109.7 a game. Austin is second at 108.6 and made up the difference with the biggest week anybody's had.

The variable is Austin's bench, which stranded roughly 50 unused points last week and grades out ninth in depth going forward. Do that again in a game where the projection gap is five points and it stops being a funny footnote.

Whoever wins this is 2-0 with a résumé. Set your lineups. Preferably with the good players in them.