I get the argument, but the estimated savings are bogus.<p>All the images on the mobile version of apple.com add up to a grand total of 500KB. If reducing their sizes by ~40% (that is, 200KB) would save 2.25s, it means the whole page (just over 3.5MB) currently takes 40 seconds to load. But obviously nobody in the developed world is waiting 40 seconds for apple.com to load. The real savings are probably somewhere between 0.05s and 0.5s depending on network conditions. Not insignificant, but much less than what is promised.<p>If you really wanted to minimize the time it takes to load apple.com, you should start with the 10 scripts, 7 stylesheets, and 9 webfonts that together make up over 80% of the page size and consume a considerable amount of resources to parse and execute. But the benchmark doesn't tell you that. It's just a checklist of micro-optimizations that doesn't even start with realistic assumptions.