The rating says how good. The identity says what kind.
A number alone flattens the thing that matters. An elite draft-and-develop front office and an elite win-now trader can land at the same tier and be nothing alike: one hoards picks and pays its homegrown core, the other ships picks and buys finished stars. The engine resolves three identity axes from the transaction record, aggressive versus patient, loyal versus ruthless, develops versus buys, and reports the fingerprint alongside the rating. And because the owner sets the budget the GM spends, it re-reads his choices against the room he was actually in.
Two front offices, same tier, nothing alike.
A cold GM search should not return one ranked list. It should return a shape. The engine reports the overall front-office rating and, beside it, a three-axis identity read from the transaction record, so two GMs at the same number are told apart by what kind of builder each one is. Two composite front offices, both rated 87.
Identical rating, opposite fingerprints. One is a patient, loyal developer; the other an aggressive, ruthless buyer. A single number would call them interchangeable and be useless the moment you needed to hire one. The engine answers both questions: how good, and good at what.
Illustrative engine read on the real rating-plus-identity structure (three axes from the transaction record). Composite front offices, demonstration figures.
A GM who will not spend may be a GM whose owner will not let him.
The GM controls allocation and structure. The owner controls budget and willingness. Judging the GM without separating the two blames him for a ceiling he did not set. The engine runs an owner-adjustment that re-reads his spend and his navigation against the owner he actually worked for. Composite GM under a tight owner.
On the surface he looks cheap and passive. The owner-adjustment reveals a hard budget ceiling and a no-luxury-tax mandate: he was not unwilling to spend, he was not allowed to. His spend and navigation grades rise once the owner's constraint is stripped out, while the owner-neutral pillars do not move. The engine grades the GM against the room he was in, and reads the owner separately.
Illustrative engine read on the real owner-adjustment structure (re-reads spend and navigation against the owner's budget and willingness). Composite GM and owner, demonstration figures.
The fingerprint is not a label. It is a fit.
Identity is only good or bad against a situation. A patient hoarder is perfect for a rebuild and wrong for a closing title window; an aggressive buyer is the reverse. The engine reads the fingerprint against the team's actual window, so a great GM in the wrong seat reads as the mismatch he is. Same aggressive-buyer identity, two teams.
The exact same 87-rated aggressive buyer is a perfect hire for one team and a slow-motion mistake for the other, and only the identity-against-window read tells them apart. A rating alone would greenlight both. Hire the fingerprint that fits the window, not just the highest number.
Illustrative engine read on the real identity-to-window fit structure. Composite GM and team situations, demonstration figures.
A front office is not a single point on a leaderboard. Two GMs at the same tier can be opposite builders, and which one you want depends entirely on the window you are in and the owner you have. The engine reports the rating and the fingerprint together, reads the fingerprint against the team's situation, and separates the GM's own choices from the budget the owner set above him. How good is the easy question. What kind, for which window, under which owner, is the one that actually decides a hire. A number hides all three. The shape shows them.
Hire the shape, not just the number.
Pro GM rates the front office, reads its identity, matches it to the window, and separates the GM from the owner.