The same man, spelled twice

137 matches were locked out of our archive because history spells footballers' names more than one way. Teaching the gate about names — without teaching it to guess — took more care than it sounds.

Andrei Arshavin and Andrey Arshavin are the same man. Pierre Aubameyang and Pierre-Emerick Aubameyang are the same man. Leônidas and Leônidas da Silva are the same man — the greatest Brazilian of his generation, and for weeks our archive refused to publish his finest match because of it.

The lockout

Our verification gate publishes a match sheet only when the page's scorers agree with an independent goal record. It compared names strictly — and strictness has a cost. When we finally measured that cost across every cached page in the archive, 137 refused matches turned out to be spelling variants of the same player: transliteration differences (Valeri and Valery Karpin), name-order conventions, short forms against full forms (Griezmann; Antoine Griezmann), compilers' styles from different decades. Every one of the 137 was re-verified against the independent record before publishing: all 137 scores matched, and 128 of 137 matched on the recorded minute of every goal.

Tolerance without guessing

The dangerous way to fix this is fuzzy matching — accept anything close enough. Archives are full of fathers and sons, brothers and namesakes, and close enoughis how a database invents a goalscorer. So the tolerance is narrow by construction: a shorter name matches a longer one only when all of it appears in the longer name anchored on the surname; a lone name may claim only the true short form of a Portuguese-style particle surname; and a single-letter transliteration swap counts only when the entire rest of the name already agrees. Above all: the comparison happens inside one match's own goal record, and if two different players in that match could both fit the name, the gate refuses exactly as it always did.

The measured result of the change: 137 matches gained, zero previously published facts altered, and the refusal ledger — which we keep, by reason, for every page we decline — shrank by precisely the one category this class explains. The Battle of Bordeaux, the Disaster of Sweden and Spain's twelve-goal night all serve because of it.