Why the same season is worth different money at 21 and 33
Two players produce identical output. The model prices them differently, on purpose, and the curve that does it is on the methodology page rather than buried in a coefficient.
A twenty-one-year-old and a thirty-three-year-old have the same season. Same minutes, same output, same level of opposition. Any model that prices them the same is wrong, and any model that prices them differently owes you an explanation of by how much and why.
What is being priced
A transfer fee is not a payment for last season. It is a payment for the seasons the buying club expects to get, plus whatever the player can be sold on for afterwards. The younger player carries most of a career and a resale value. The older one carries a contract and, often, a free transfer at the end of it.
So the age adjustment is not a bias correction bolted onto a performance model. It is part of the thing being valued. Removing it would not make the model more objective; it would make it answer a question nobody asked.
Why the curve is visible
The multiplier applied at each age is drawn on the methodology page and on the landing page. That is deliberate. An age adjustment sitting inside a fitted model as an unlabelled coefficient is the sort of thing that quietly drifts, and nobody notices because nobody can see it. Drawn as a curve, a wrong shape is obvious at a glance — including to a reader who thinks it is wrong and wants to say so.
The peak sits where the market's own behaviour puts it, not where a scouting cliché does, and the decline on the right is gentler than the drop-off people assume. Both of those are checkable claims, which is the point of drawing them.