I think we agree for the most part and there was a misunderstanding regarding interpreting frequency response.
If two headphones have the exact same frequency response at your own eardrum, they will sound exactly the same. All relevant factors such as resonances, angle of incidence etc would have a measurable effect on frequency response, if accurately measured in your own eardrums. This is a hypothetical scenario and does not reflect the reality of headphone measurements.
You cannot look at a 2d frequency response graph from a gras and know precisely how headphones will sound to your ears. I should stated this clearly earlier! Furthermore, there is information characterized by the frequency response that cannot be understood by anyone by simply looking at a 2d graph, for example resonances. Frequency response is great for providing a general idea of how a headphone sounds and providing a baseline for eq, but looking at a graph is a limited tool. I think we both agree on this.
This actually is my main criticism of Amir's headphone reviews. I believe he should listen before he measures. He measures headphones and evaluates them, with frequency response being the most important measurement. To me, it seems like he overvalues his analysis of frequency response and it biases his listening. For example Amir's measurements of a poorly sealed Ananda and its subsequent bad looking frequency response graph seem to bias his listening in a similar way that a low price biases crinacle or resolve when they listen to a cheap headphone. It is a bit unfair to compare Amir to these reviewers - he has no subjective evaluation of macrodynamics or detail - but the comparison is useful for illustrating my criticism. Funnily enough, my criticism doesn't apply to the stealth review as he listened first. My criticism in the scheme of things is pretty minor and I like Amir's headphone reviews - I think he does a much, much better job than most reviewers.