Why friendly matches distort a rating system, and how weightings answer it

The assumption a rating quietly makes

Every rating system assumes that a result reflects an honest attempt by both sides to win with the strongest available personnel. A match arranged for preparation breaks that assumption, because one side may be experimenting with selection while the other treats it as a genuine test. The scoreline that emerges is real, but it is evidence about a situation that will never recur in the competition the rating is meant to describe.

Feeding that result into the ledger at full value therefore imports information about squad depth into a measure of first-choice strength. The distortion is largest exactly where it is least welcome, which is at the top of the table where selections vary most freely.

Unequal incentives on the same pitch

The two sides in a preparation match frequently have different objectives, and asymmetric motivation produces results that mislead in a predictable direction. A side treating the fixture as a showcase will field its best and press for a decisive win, while the opponent rotates and manages workload. Nothing about that arrangement is improper, yet the resulting margin says far more about priorities than about the relative quality of the two squads.

Home advantage compounds it, since preparation matches are often arranged on favourable terms as part of the commercial agreement. A rating system cannot see any of this, and it records the outcome exactly as it would record a decisive competitive result.

Weighting as the standard remedy

The usual answer is a multiplier that reduces the influence of low-stakes matches, so they still count but cannot dominate a rating. A small multiplier preserves some information, which matters for sides whose competitive calendar is otherwise sparse and who would otherwise stop moving entirely. Choosing the size of that multiplier is a judgement about how much of a preparation result is genuine signal, and no measurement can settle it.

Set too high, the multiplier reintroduces the distortion, and set too low it strands whole regions with almost no rating-relevant fixtures. Most systems land on a value that keeps friendlies visible but incapable of moving a side more than marginally within a cycle.

The case for excluding them entirely

Removing preparation matches altogether produces a cleaner measure of competitive strength and eliminates any incentive to schedule opponents for rating gain. The cost is a large information gap, particularly for sides between qualification cycles whose ratings would then freeze for extended periods. Frozen ratings age badly, since squads change and a side seeded on a two-year-old assessment may bear little resemblance to the one that earned it.

Exclusion also concentrates all rating movement into tournaments, which increases volatility at exactly the moments when stability is most valuable. The trade is between a cleaner measure that updates rarely and a noisier one that stays current, and neither dominates the other.

Designing for the behaviour you want

A weighting scheme is not only a statistical correction but an incentive structure, because associations schedule with the rules in mind. If low-stakes wins are cheap, the rational response is to seek meaningful fixtures, which is the behaviour the sport benefits from. If they are valuable, the rational response is to arrange comfortable ones, which degrades both the calendar and the rating simultaneously.

Good design therefore asks what scheduling pattern the rules encourage before asking whether the arithmetic is elegant. That question is answerable in advance, unlike the deeper problem of measuring effort, which no formula can observe directly.