Is there not also the possibility where the baseline of <X> is rather low, thus allowing for a comparatively "highest ratio" of <X> to <Y> without implying a low <Y>?

Not saying that's necessarily the case here, I don't follow the blog either, just pointing out another interpretation.