The ranking system explains less than its confident users assume

A formula, not a measurement

A ranking assigns points under fixed rules covering which matches count, how much each is worth and how quickly older results decay. Every one of those rules is a choice, and changing any of them reorders the table without anything changing on any pitch. That makes the order an output of a set of decisions rather than a reading taken from the world.

It can still be useful, in the same way that any index is useful once its construction is understood by the people quoting it. The difficulty is that it is quoted overwhelmingly by people who have not examined the construction.

Unequal fixture opportunities

National teams do not choose their opponents freely, since confederation structures determine the majority of competitive matches available. A team in a strong region accumulates results against strong opposition while a team in a weaker one has fewer chances to demonstrate anything. Formulas attempt to correct for this by weighting opponents, and the weights are themselves derived from the same rankings being computed.

That circularity is unavoidable in any system with no external measure of strength to anchor against. It means regional biases propagate rather than being corrected, and no amount of formula refinement removes the underlying problem.

Friendlies and the incentive question

Non-competitive fixtures carry lower weight, which changes the incentive to arrange them and against whom. A federation aware of the weighting may prefer opponents whose results are less likely to damage a position it wants to protect. That is a rational response to a published formula and it degrades the information the formula was meant to capture.

Any metric that becomes a target for the people it measures loses some of its meaning, which is a general pattern rather than a specific failing here. The effect is difficult to quantify and impossible to dismiss. A formula that changes behaviour is no longer a neutral observer of the thing it was built to describe.

Why small gaps mean nothing

Two teams separated by a few places have point totals close enough that a single fixture with a different opponent could reverse them. Treating that separation as a statement about relative strength attributes precision to a difference that is well inside the noise of the process. The gap becomes informative only when it is large, and the discussion is almost always about the small ones.

This is the same error as reading a difference in a batting average without asking whether it exceeds the variation the number contains. Precision in presentation and precision in meaning are entirely separate properties.

What the table is good for

The order is a reasonable coarse summary of which teams have been winning against decent opposition over a recent period. It is also a seeding instrument, and for that purpose it needs only to be defensible and consistent rather than exactly right. It is a poor instrument for settling arguments about which of two teams is better, which is what it is mostly used for.

A tool that performs its designed function adequately can still be misleading in the use it actually receives. The gap between design and use is where the confident quoting does its damage. A seeding instrument does not have to be right about who is better, and a debate about who is better needs an instrument that does.