You are not playing the game.
You are playing everyone else.
Drop a screenshot. The Doctor reads your fifteen, prices every fixture off the betting market, and plays out the next ten gameweeks ten thousand times against the squad everybody else owns — so you can see how far ahead you are, and how hard it can swing.
What the screenshot read as
Correcting a slot and re-diagnosing costs nothing and reads no screenshot — it posts the fifteen player ids straight to the model.
Each tile shows the club, the price, and the bar is how much of the field effectively owns him — captains counted twice, which is why it can pass 100%. The figure on the strap is how much of the field you beat every time he returns.
Share it
Mint a link to a card of this diagnosis — the thing to paste into a rate-my-team thread instead of a screenshot.
How this works
What is “the template”?
The most-owned legal fifteen: take everyone's ownership, buy the squad that maximises it under the real budget, position and three-per-club rules, and that is the team the field is effectively playing. Nobody owns exactly it — it is the centre of gravity all those squads sit around, and the thing you are actually beating or losing to every week.
What do the two numbers mean?
The first is how many points a week your eleven is worth over the template's. The second is how far a single week can land from the pack in either direction — a squad of pure template players moves with everyone, so it barely swings; a squad of differentials can jump twenty points clear or twenty points behind on the same Saturday.
Read together they say what kind of season you have signed up for. A small edge and a small swing means you will finish near where you started. A large swing means you will not, in one direction or the other.
Both are for a single gameweek, and both come from ten thousand joint simulations rather than from a formula — players who share a fixture score together, and a model that ignored that would understate the swing badly.If you know the portfolio language: these are alpha and tracking error.
Why does the fan on the chart get wider?
Because every gameweek adds its own swing on top of the last one, and those swings do not cancel. Eight weeks of “anything within ten points” is not “anything within ten points” at the end — it is roughly three times that either way. The line in the middle is the edge compounding; the widening around it is the risk compounding, and the second one grows faster than most managers expect.
The chart plots the difference against the template rather than two running point totals, because two lines both climbing past four hundred points would hide a thirty-point gap between them completely.
Why is my differential's ownership over 100%?
That figure is effective ownership, which counts captaincy twice. A premium captained by half the field is effectively owned by more than everyone, and that is exactly the exposure that decides ranks — so the bar clamps at 100% but the number does not.
What does “hold” mean on a chip?
That several gameweeks are inside the noise of each other, so no week is distinguishable as the best one. It is a real answer, not a missing one: the right move is to keep the chip until a week separates.
Where a week is called, the figure beside it is the extra points on top of what your squad was going to score anyway — a chip is only ever worth the difference it makes, never the total it appears on.
Is the screenshot reading reliable?
Mostly, and where it is not it says so. A model reads the names off the image; a deterministic matcher — no model, no guessing — resolves those names to real players and flags every slot it is not sure about, including ones where the shirt colour disagreed with the name. Those are the slots you are asked to check.
Correcting one and re-diagnosing reads no screenshot at all: it sends the fifteen player ids, which is both free and exact.
Where do the fixture ratings come from?
Club strength comes from BayesFC's Bayesian ratings, fitted to the betting market rather than to results — so they say what the bookmakers think of a club this week, not what its results averaged out to over a season already gone. Every scoreline in the simulation is drawn from those ratings, and their pull fades the further ahead the week is, because nobody knows a GW38 line-up in December.
If those ratings are ever unavailable the diagnosis still runs, with every fixture priced level, and says so in a note at the top. It never quietly hands you neutral numbers as though they were real ones.