Position scarcity and why like-for-like comparison breaks

Different jobs produce different evidence

A competitor whose role is to create scoring chances leaves a very different statistical trace from one whose role is to prevent them. The first accumulates visible, countable events, while the second is measured largely by things that did not happen while they were present. Any direct comparison of their raw outputs is therefore meaningless, because the two are not producing the same kind of quantity at all.

This is not a data problem that better collection would fix, since the asymmetry follows from the structure of the roles themselves. A ranking that mixes roles must first convert both kinds of evidence into a common unit, and that conversion is where the arguments begin.

Positional baselines as the standard device

The usual approach compares each competitor against a typical performer in the same role rather than against the field as a whole. That produces a figure expressing how far above or below the positional norm somebody sits, which is at least dimensionally comparable across roles. The method depends entirely on defining the norm, and a norm computed from a shallow pool will differ from one computed from a deep one.

It also assumes roles are stable enough to have norms, which is increasingly questionable in sports where tactical systems blur traditional boundaries. When a role is being redefined, the baseline lags behind the reality and every competitor in that role is mismeasured in the same direction.

Scarcity and the value it creates

A performance has a market value that depends on how many others could plausibly deliver something similar, which is a question about supply. Where a role is thinly populated, an ordinary performer in it is harder to replace than a slightly better performer in a crowded role. Ranking systems that incorporate scarcity therefore elevate competitors whose absolute output is modest but whose substitutes are genuinely poor.

This feels wrong to readers who understand a ranking as a statement about quality rather than about replaceability, and both readings are legitimate. Saying which one the table intends removes an argument that otherwise repeats endlessly without either side changing position.

The interaction with team context

A role only exists inside a system, and the same nominal position can carry entirely different responsibilities under different tactical arrangements. A competitor asked to cover a large area defensively will produce numbers that look poor beside somebody in a more protected version of the role. Adjusting for system requires knowing the instructions, which are not observable, so analysts infer them from behaviour and then use that inference as a correction.

The circularity is uncomfortable, because the correction is derived from the same performance it is meant to contextualise. Acknowledging the limitation is more useful than adding another layer of adjustment on top of an already unstable estimate.

When cross-role ranking is worth attempting

For selection and squad building, cross-role comparison is unavoidable, because a finite budget must be allocated across every position at once. For public argument about who is better, it is far less necessary, and the insistence on a single ordering usually generates more heat than insight. Ranking within roles produces tables that readers trust, because the comparison is between competitors doing recognisably the same job.

Those within-role tables can then be combined for the purposes that genuinely require it, with the conversion assumptions stated openly. The sequence matters, since converting first and comparing afterwards hides exactly the assumptions a reader most needs to see.