I agree that taste plays a major role in any evaluation of an interpretation; so do musical education and experience, which tend to reshape your taste over time.
When you say, you like or dislike an interpretation, it’d be unbearably arrogant to tell you that you are wrong to do so, but that doesn’t exclude a discussion of the skill of the player, the consistency of his approach and its variation from other interpretations and the likely intentions of the composer.
The Ballade No. 1 in G minor, Op. 23 is a fine piece to teach your ears to hear all of that and more.
It took Chopin four years to complete it; the final work was done when he arrived in Paris and was admitted into society.
The ballad is influenced by one of Mickiewicz’ poems, Konrad Wallenrod, and was dedicated to Monsieur le Baron de Stockhausen.
“So what?”, you might ask. Well, if you know the poem, for instance, you realize immediately why Chopin used alternating keys and what he wanted to express by it; you also understand that Chopin “narrates” a developing story that leads to an explosive conclusion.
If you have an idea about the intentions of a great composer, you have a much easier time to understand how he translated them into notation and why he chose, for instance, in m.216 and m. 218 Neapolitan chords held with an accented bass C.
Just think about the differences you’d hear, if he had chosen an iv chord …
The better you understand the accents, the melody, and the harmonic structure Chopin chose, the more you will get his idea and you will hear more and more differences among various interpretations.
An interpretation isn’t better because it follows Chopin’s intentions more closely, but the pianist should have a reason to deviate and his intentions should be consistently heard and create a harmonious or integral whole. Otherwise, he is second rate.
And btw., Chopin’s and Beethoven’s sonatas almost always tell you something about the pianist’s hands: he or she might play incredibly well, but only large hands will be able to play the entire score. 