So, a caveat here: I don't know One Piece. I've picked up a bit by osmosis, but I never have and never will actually watch the show, as highly episodic weekly shows that rack up hundreds of episodes are very much not my thing.
That doesn't necessarily follow. The overall threat level isn't dictated by the outliers, particularly in the context of a large sample set. If the potential opposition is a giant pile of mooks with one real threat that completely outclasses you buried in it it's a lot less dangerous than a potential opposition of a giant pile of mooks liberally seasoned with real threats, even if said real threats are comparable instead of far better.
This would be more convincing, though both the randomness of the sea beast and the extent to which it's local wildlife have been called into question. Which gets back to the influence of the threat levels of the outliers. An ocean with the leviathan world serpent which will eventually be involved in the apocalypse and kills everything it comes across, and also a bunch of small filter feeding fish is generally a whole lot dangerous than an ocean with a bunch of small filter feeding fish, providing the food baseline for a whole ecological level of massive and dangerous sharks, sea serpents, kraken, etc., even if said sharks serpents and kraken pose roughly the same threat to the leviathan world serpent as those small feeder fish.