Power rankings

1. The math: record + deserved wins

Half the board is your actual record. The other half is deserved wins, which asks what your scores were worth against the whole league instead of against the one team the schedule handed you.

Every week we take your score and ask a simple question: how many of the other eleven teams would you have beaten? Drop 140 in a week where nine teams scored less, and you banked nine-elevenths of a win — about 0.82. Do that every week and it adds up. That total is your deserved wins, and it is the whole ranking.

No opinions, no projections, no roster ratings. board = ½ × wins + ½ × deserved wins. Win games, and score more than more people. That is it.

Deserved wins on its own is schedule-blind on purpose — but used alone it said a 1-2 team was fourth and a 3-0 team fifth, because a blowout banks the same single win as a squeaker. Keeping the W-L column in the board fixes that without pretending the schedule is fair.

Schedule luck is your real wins minus your deserved wins. Positive means the schedule was kind to you. Negative means you have been getting mugged. Across the league it always adds up to exactly zero — every bit of good luck came out of somebody else's hide.

“vs the league” shows it week by week. “9-2” means your score beat nine of the other eleven that week.

2. Two tiebreaks

Deserved wins is a decimal, so two teams can land a hair apart on nothing. When they are inside one opponent of each other, the better record breaks the tie, then points for.

Then head-to-head, with a leash: if you beat somebody this week and they are within three spots of you, you go above them. Further apart than that and the season's body of work wins — beating the best team once does not make a last-place team good.

3. The commissioner's pass

The math cannot see who is still standing. So the commissioner may move any team up to three spots, and only for one reason: roster news — a starter lost, a starter back, or a score that was wrecked by somebody getting hurt in the first quarter.

Everything else is a story, not a reason. Blown leads, comebacks, walk-offs, leaving 40 points on your bench — all great copy, all worth reading about, none of it moves a rank. The scoreboard already counted it once.

And a team never drops below one that is worse on both record and deserved wins. Losing your quarterback is the story of your season; it is not evidence that the 0-2 teams under you are suddenly better.

Where the commissioner disagreed with the math, the row says so. Most weeks, most teams do not move — that is the system working, not the commissioner asleep.

Three terms

Season

FFWrapped raw = (avg weekly pts × 6 + (best week + worst week) × 2 + win% × 400) ÷ 10, regular season through W, with win% = ½ actual + ½ all-play (schedule luck halved)

Recent

last 3 weeks weighted 70/20/10: 65% × (PF − league mean that week) ÷ 5 + 35% × result, where result = ½ actual W/L (W 10, T 5, L 1) + ½ all-play that week (1 + 9 × share of teams outscored), +0.5 week high, −0.5 week low — schedule luck halved, as in the season term

Roster

40% depth + 40% leverage + 20% bench. depth = FantasyPros weekly positional ECR of every rostered player → value 100 ÷ (1 + ((ECR−1) ÷ half)²), depth-weighted per corps, each corps scaled 1–10 across the league (mean 5.5, best 10), corps combined by position weight. leverage = the best lineup the roster can start next week, in this league's points (FantasyPros projection + first-down/yardage bonuses), each player × probability of playing this week (FantasyPros' number, else OUT/IR/PUP/Suspended 0, Doubtful 0.25, Questionable 0.75), every slot floored at replacement level — the average of the best 5 still available at that position in the newest free-agent snapshot, taken after the Wednesday 07:00Z FAAB run so the names left are genuinely free (which also lets the week's waiver results move the board), valued the same way as any rostered player — because a manager with an injured starter streams a replacement rather than fielding nobody; the floored slot is listed as '(waiver replacement)' and the rostered player stays eligible for a FLEX behind it. Depth and bench stay roster-only: renting a starter off the wire is not the same as owning one. bench = the best 4 non-starting RB/WR/TE projected the same way (a backup QB/K/DST cannot start here), blended with the RB/WR/TE points the bench has actually scored per week (the actual half weighted min(1, W/4)). Injury status: FantasyPros when they list the player, else ESPN's own injuryStatus

Scaling

every term is min-max scaled 0–1 across the league before weighting; ties → season raw → name

Weights by week

season = min(0.45 + 0.04W, 0.80) × min(1, W/4); recent = max(0.20 − 0.01W, 0.06); roster = remainder

Roster term

Mix

depth 40% · leverage 40% · bench 20%

Corps weights

QB 14.5% · RB 34.0% · WR 34.0% · TE 11.5% · DST 6.0%

Depth weights

QB: 1.0/0.25 · RB: 1.0/1.0/0.8/0.5/0.4/0.25/0.1 · WR: 1.0/1.0/0.8/0.5/0.4/0.25/0.1 · TE: 1.0/0.25 · DST: 1.0/0.1

ECR half-rank

QB 12 · RB 24 · WR 30 · TE 8 · DST 8 — the positional rank at which a player is worth half of the #1

Bench

best 4 non-starting RB/WR/TE

Availability

ranked this week: this-week ECR × probability of playing if Questionable/Doubtful. unranked (injured/suspended): preseason ECR × OUT 0.6, PUP 0.5, IR 0.4, SUSPENDED 0.5, DOUBTFUL 0.5, QUESTIONABLE 0.8

This league's points

Projections

FantasyPros PPR projection + first downs (yards ÷ yards-per-first-down: pass 21.1, rush 17.6, rec 21.1, measured on nflverse 2023–24) × the rule value + expected 300/400-yd passing and 100/200-yd rushing/receiving games × their bonuses; 50+ yd TDs not projected

Editorial pass

the published Commissioner edition may move any team at most 3 spots from this board; every move must be justified by that week's verified numbers; the site prints 'math #n · moved ±k' beside each moved team

Not in the formula

ESPN's currentProjectedRank (shown as a reference); FantasyPros League Analyzer (not in their API); any external edition