The Cheats' Fines Were Cast into Statues and Set Beside the Entry TunnelThe Zanes at Olympia · 388 BCE, Elis
388 BCE (98th Olympiad)OlympiaPausanias 5.21 · Young 1984Months before the games, Elis sent heralds out to proclaim the sacred truce (ekecheiria) — which did not stop the wars, only guaranteed that those travelling to the festival would not be attacked on the way. The judges were chosen by Elis, and competitors had to arrive a month early to train and be vetted there. The whole apparatus solved exactly one problem: why should city-states currently at war with each other accept the same result.
In 388 BCE the Thessalian boxer Eupolus bought off three opponents. The fine was not taken into the treasury; it was earmarked for six bronze statues of Zeus (the Zanes), set along the vaulted passage from the judges' stand to the stadium, their bases inscribed with the cheats' names and cities — every competitor thereafter had to walk past them on his way in. Pausanias, travelling in the second century CE, could still count sixteen. As for rewards: the victor formally received only an olive wreath, and a great deal more once home. Athens, from Solon onward, paid an Olympic victor 500 drachmas — roughly the price of 500 sheep.
Without the truce and the cross-polis judiciary there is no panhellenic festival at all, and no Greek legacy of "results comparable under one set of rules". The dispute is over its character. The Victorians made the ancient games the fountainhead of amateurism; David Young, The Olympic Myth of Greek Amateur Athletics (1984), used the prize money and the professional athletes to show that this was a nineteenth-century invention — "amateur" was a rule for keeping mechanics and labourers out. Nigel Spivey, The Ancient Olympics (2004), adds the other cut: it was first of all a religious festival and quasi-military contest.
For any cross-institutional leaderboard or evaluation, the cost was never in the scoring. It is in getting mutually distrustful parties to agree on who scores and how cheating gets displayed — exactly where the credibility crisis of model benchmarks is stuck.