The Peloponnesian fleet—fearing the arrival of the Athenian reinforcements—even withdraws under cover of night as far as Corinth (except for the ships from Leukas). 'And the Athenians from Crete, with the 20 ships that should have arrived for Phormio before the battle, arrived not mud later than the retreat of the ships at Naupactus' (92.7).With this last sentence the author has followed the fragmentation of the whole undertaking to its final individual action. What is emphasized is that the squadron arrives shortly after the battle and thus could 'almost' have joined the attack. In literary terms this statement is a reference to 85.6-86.1 and the observation at 86.5 that the two fleets lay at anchor opposite each other for six or seven days without engaging. At the end Thucydides thus once again underscores the difference between 'what should have happened and 'what really happened-an aspect of the whole course of events which has been essential since Knemos' first move.It seems useful, in the light of this distinction, to look again at the second sea battle in which both Phormio's plan and the Peloponnesian commanders' probability-calculations (87.7) are brought to nothing by the actual course of events. De Romilly' interpretation runs thus: a slight element of chance (91.3) restores the 'normal state of affairs (i.e. Athens' experience-based naval superiority and Peloponnesian incapacity to fight at sea), thus justifying after the fact Phormio's plan, which had been based upon 'normal' battle conditions: ‘Thucydides...presents the final result more as necessary than as accidental’.