The Tape

01 · THE HONEST STANDINGSTwo undefeated teams, and they have not met

Record next to all-play — your record against every team, every week. Luck is real wins minus the wins your scores earned.

Alex and Andrew are both 3–0 and have not played each other. Alex has faced Logan, Louis and Michael; Andrew has faced Jerome, Pat and Ernie. Neither has beaten the other at anything except the all-play table, and that table now reads 26–1 to 23–4. They meet in Week 5.

Behind them, five teams are 1–2 and two of them have a winning all-play record. Michael and Jerome have both beaten more teams than they have lost to; Brent is a coin flip at 13–14. All three are two games under. That is not a slump, it is a schedule. Logan is the reverse: 2–1 on a 5–22 all-play, the widest gap between record and performance in the league.

02 · RECEIPTSFour of five, and that’s not the story

What last week’s preview said, what had already changed by kickoff, and what happened. Red is a call the set lineups had going the other way.

Four of five sounds like skill. It isn’t. These calls came off Yahoo projections for lineups as they stood Monday night, pre-waivers. By kickoff the favorite had flipped in two of them — Logan over Louis was called by ~20 with the set lineups favoring Louis, and Logan won by 30.30 only because Louis’s starters came in 28.55 under, the worst in the league. The average lineup moved 11.88 points after those calls went out, against 14.28 points of player variance. Nearly half the error was roster churn, not football.

The one miss is the best thing on this page. Alex went into the week with one quarterback, Caleb Williams, who landed on IR. He went and got Bryce Young, started him for 14.64, and beat the team this page picked to finally end his streak. The call was right about everything except whether Alex would do something about it.

03 · WEEK OVER WEEKTwo teams moved six places. One fell seven.

Louis has gone 165.06, 121.32, 86.66 — down roughly forty points a week, in a straight line, for three straight weeks. Only Alex has also fallen every week, and he is falling from 168.06. Pat is the mirror image: 99.46, 116.54, 126.20, climbing every week and still 0–3.

04 · THE COACHING TAXThe best lineup and the worst were set by the two winners

Points left on the bench: your optimal lineup minus what you actually started.

Andrew left 1.82 on the bench and it was Patrick Mahomes. His entire coaching error for the week was starting Jalen Hurts over a quarterback he has now benched for three straight weeks — 53.14 in the first two, 16.94 in this one. He posted the highest score of the week anyway. That is what a clean lineup looks like: the only thing you got wrong was the thing you have been getting wrong all year.

At the other end, Logan left 26.52 — Burrow 23.08 and Luther Burden 19.80 both on the bench — and won by thirty. Louis left 20.30 out of a 106.96 ceiling, which means even a perfect lineup would still have been the lowest score of the week.

05 · THE BIG ONE, REVISITEDRound 3 was never a trap

Average Week 3 points by the draft round a starter came from.

Two threads died this week. Round 3 is a trap was Confirmed after two weeks at 13.35 and 10.56 — it returned 19.41 and finished third. Round 2 is the best round was the other half of the same story, and Round 2 finished fifth while Round 1 put up 26.66. Bijan Robinson 42.80, Jahmyr Gibbs 44.90 and Christian McCaffrey 23.60 are all Round 1, and all three went in the first four picks.

The draft-slot correlation, which opened the season at −0.85 and looked like the scariest number in the league, came in at −0.33 this week: slots 1–5 averaged 138.20, slots 6–10 averaged 132.28. Five and a half points. It was noise.

06 · CORRELATIONBuffalo stopped scoring and took Louis with it

How much of each Week 3 score came from one NFL offense.

Louis’s Buffalo share has fallen every week and it is not because he diversified. 46.7% and a win, 68.1% and a loss, now 30.4% of the worst score anyone has posted. The stack shrank because Buffalo stopped producing: Kincaid 3.80 and Shakir 1.60 out of two starting slots. Last week this page said the single point of failure was everything that isn’t Josh Allen. This week Allen was 20.96 and it still wasn’t close.

Alex is the one to watch for a different reason. Four 49ers, and two of them are the kicker and the defense — the pair that cannot both pay off in the same game. They returned 9.00 between them out of two roster spots. He won by 13.48 anyway.

07 · WEEK 3 PERFORMANCESThe best of it, and the worst

Four of the seven starts on this list belonged to two managers. Louis had three of them and lost by thirty; Michael had the quarterback and lost by thirteen with the second-best score of the loser’s bracket.

08 · POSITIONAL EDGEWhere each team wins, season to date

Points above or below the league average at each slot, Weeks 1–3 combined.

Alex’s season is still two running backs and a flex. +75.8 at RB and +51.0 at FLEX against the worst receiving corps in the league at −42.1. Three weeks in, the hole has not cost him a game. Andrew is the only other team with a flex above +40, which is most of the difference between the two 3–0s and everyone else.

Logan’s −52.5 at running back is the largest single-position deficit in the league, and his two positives are tight end and kicker. Louis has the best quarterback in the league — Allen at +39.6 — and has still outscored only Logan and Pat. Michael has the best defense at +25.7 and the worst quarterback at −31.3, which is a roster you could fix with one trade nobody is making.

09 · KEEPERSThree weeks of return on the keep

Javonte Williams at a round-8 price is the best keeper in the league and it is not close. Jonathan Taylor at round 2 is the worst — Alex is paying a second-rounder for 10.70 and winning anyway, which is its own kind of insult to the rest of the field.

10 · AVAILABILITYWho’s missing for Week 4

Logan started Achane on Sunday for 2.20 and he is on injured reserve by Monday. Louis just posted the worst score anyone has managed this season and his quarterback is questionable.

11 · WEEK 4The preview

Yahoo’s projections as of Monday night, before waivers run. Treat these as a starting line, not a forecast.

Pat is 0–3 and gets the closest game on the board. Brent is 1–2 with a winning all-play and draws the hottest team in the league in a four-point game. And Logan, at 2–1 with the worst all-play record in the league, is a 27-point underdog to a 1–2 team. The schedule is about to start telling the truth.

12 · PREVIOUSLY ON EPSOFFThe season so far, one thread at a time

Every thread carries the week it opened. Nothing retires the week it resolves — the league gets one more week to argue about it.

Nothing retired this week, and that is the story. Two threads were sitting on the retirement watch — Round 3 is a trap at Confirmed and Natchez is a fraud at Reversed, both due to close after Week 3. Both flipped instead. Natchez answered the fraud charge in Week 2 with 153.06 and then answered the answer with 112.44 and a 1–8 all-play. The thread is three weeks old and further from settled than it was on day one.

13 · WHAT TO WATCHThree open questions

Does Louis stop the slide? 165.06, 121.32, 86.66. He bet the draft on Buffalo, Buffalo delivered 30.4% of an 86-point week, and Nacua and Collins are both still out. His best possible lineup on Sunday was 106.96.

Is Andrew the actual best team? Both are 3–0. Alex has the better all-play, 26–1 to 23–4, and has scored more. Andrew put up the third-highest score anyone has managed this season and set a 1.82 lineup doing it. Nothing settles it until Week 5, which is the first time they play — and they meet again in Week 14.

Will anybody trade? Twenty-five transactions, zero trades, three weeks. Five teams sit at 1–2 with rosters that don’t match their records. The Week 6 Identity Report is going to name the trades whether or not anyone makes them.