If you've been to the beach, you've probably noticed that the ocean doesn't always look particularly blue, and whether it looks blue, black, green or brown can depend on the weather, location, and viewing angle. However, if the average (let's say English-speaking) person tries to paint an ocean scene from memory, they are extremely likely to use blue paint. Blue is a convenient shorthand for a rather visually/chromatically complex stimulus (a vast, constantly moving, reflective body of water), and sometimes we conflate the shorthand with real qualities of the actual object, even when we can observe that our blue painting of the ocean doesn't resemble what our eyes really see. Essentially, bottom up perception can easily be lost in translation once our sensory organs pass the task of expressing what we've seen along to the rest of our cognitive framework.
I am fairly certain ancient Greeks were able to perceive the color blue. They traded with the Egyptians, who created spectacular blue dyes/pigments, and Homeric epics described Zeus's eyebrows as κυάνεος (kyaneos), which covered a spectrum of greys, blues, and purples. There was even a sea god named Γλαῦκος (glaukos) whose name was also a color word for shimmering blues and greens (and maybe yellows?). Perhaps they had a more nuanced familiarity with the color blue and just didn't lump kyaneos and glaukos into a shared category of blueness like we might.
Now, although I don't find the idea that ancient Greeks completely lacked a concept of blue particularly convincing - as you pointed out, the age old question of the wine-dark sea really does lead to some interesting considerations about language and perception.
Per Gladstone's examples, οἶνοψ (oînops/wine-dark) is used to describe the churning sea and imposing oxen. It really makes me wonder if the oînops is actually a Lakoff-esque reference to a looming threat. Maybe a stormy, dangerous ocean actually appeared to be a menacing dark red to the poets of the day, the same way the ocean seems blue to us in our minds even though it often isn't. Not exactly as a metaphor in the deliberate literary sense, but in a less-conscious, automatic processing kind of way. After all, in English we "see red" when we're angry, and this link between the color red and anger/aggression/danger is not limited to English. Priming studies have suggested that when participants are shown a series of visual stimuli priming for "anger," they were more likely to interpret an ambiguous target stimulus as being red (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3399410/). And a study on affect and color associations found that, across 30 countries and a variety of language families, red was consistently associated with anger. (https://journals.sagepub.com/doi/abs/10.1177/0956797620948810)
Or maybe wine-dark just refers to the tempestuous choppy surface of the sea (like wine sloshing in a glass), and not the color. It's fun to speculate, even if we can't really know.
As I'm sure you've noticed, talking about anything that touches on Sapir-Whorf concepts can tend to pull the conversation into tricky territory. Doubly so when the concept is applied to ancient civilizations, since we can really only guess, and our speculation inevitably reveals more about our own biases than it does about other languages and cultures, and how they perceive the world. Either way, it's an idea worth exploring, even if it's full of trap doors and windy cul-de-sacs.