Methodology

Every manager gets a rating. Here is how it moves.

An Elo rating per manager, folded over every game your league has ever recorded — then a separate ledger for what you actually won. They disagree on purpose, and this is the page that says why.

Two numbers

One asks how good you are. One asks what you have done.

Nearly every argument about these numbers ends the moment this distinction lands. They are not two attempts at the same measurement.

The League Rating asks

How good are you right now?

Zero-sum and self-erasing. Every point you win is taken off an opponent, nothing accumulates, and each offseason pulls everyone back toward average. It is a running estimate of present strength, not a record of what you have done.

The all-time standing asks

How much have you accomplished?

A ledger, in wins above an average manager. Every part is rate × opportunity rather than a rate alone, so a great half-season can only earn a little — there was only a little to earn. Nothing is ever given back.

The dials

What moves a rating, and by how much

The same numbers in every league. A game is weighed by the margin, by how surprising the result was, and by whether it was played in January, October or the consolation bracket.

Everyone starts at
1,500
A new manager enters here the first season they field a team, so the number always means the same thing: above it is above average, below it is below.
A regular-season game moves it up to
±20
The most one game can swing a rating before the margin and the odds are applied. Beating a heavy favourite moves it most; beating a team you were expected to beat barely moves it at all.
Playoff games count
1.25×
January is worth more than October, but not by so much that one bracket rewrites a season.
Consolation games count
0.5×
A losers-bracket blowout is still football, but it should not move a title-race rating much.
Each offseason carries over
75%
Everyone is pulled a quarter of the way back to average between seasons, so a decade-old run is not permanent and a returning manager re-centres softly.

Memory

The rating remembers about 2.4 seasons

This is the single most useful fact about it, and the one that explains most surprises. A rating that never forgot would not be a rating — it would be a trophy count with extra steps.

1,500 — averagehalf the edge gone — about 2.4 seasons1,742peak1,682+1 yr1,636+2 yr1,602+3 yr1,577+4 yr1,557+5 yr1,543+6 yr
A peak of 1,742 left alone, with no games to move it. Solve 0.75n = 0.5 and the rating turns out to remember about 2.4 seasons. Anything older than that is more than half forgotten — which is exactly why a current rating cannot carry a career, and why the all-time standing reads your peak instead.

The trophy

What a championship is worth, priced twice

The biggest single reason a rating and an all-time rank disagree. Both prices are deliberate: a rating should not treat a four-game bracket as proof of strength, and a league that plays all year for one trophy obviously should.

Career value pays

+11.00

Wins above average, permanently, plus that season's playoff berth. It never regresses and never expires.

The rating pays

+1.6 to 3.2

A title run typically moves a rating 4080 points, and only the part that sets a new peak reaches your career value at all. A quarter of the rest is gone by next season.

The other number

Career value is a different measurement

The all-time standing is a ledger in wins above an average manager, and nothing on it ever expires. It reads your peak rating, never your current one — which is why a manager can top it while sitting mid-table on form.

How career value is worked out

Ratings questions

What people ask when they see their own row

Join the waitlist

Numbers your league can actually argue about

Every season you have ever played, folded into one rating and one all-time table — with the arithmetic published, so the debate is about the football.

Explore the demo league

Free while in beta · invite-based signup · we'll send yours.

Invite-only beta · free