How is my statement "moving the goalpost"? I'm not moving any goalpost; I never claimed at any point that JS microbenchmarks were a good indicator of overall browser performance.
At a time when JS execution was a larger part of browser workloads (and the implementation techniques of JS engines were extremely simple compared to what was known to be possible for dynamic languages) JS microbenchmarks were a better proxy for whole browser peformance, although they weren't great at that point either. Today, all popular browsers have relatively decent JS engines and differences on microbenchmarks just don't reflect total browser performance as much as they did in the past.
Surely JS execution is a larger part than ever of browser workloads, or are you referring to things like CSS3 which move some of the rich-webapp work out of javascript?
At a time when JS execution was a larger part of browser workloads (and the implementation techniques of JS engines were extremely simple compared to what was known to be possible for dynamic languages) JS microbenchmarks were a better proxy for whole browser peformance, although they weren't great at that point either. Today, all popular browsers have relatively decent JS engines and differences on microbenchmarks just don't reflect total browser performance as much as they did in the past.