But we are talking about a difference so negligible it hardly matters. Unless you are viewing the page, or any other page for that matter, on a cellphone from the 90s with a network connection to match its probably only going to be 3 seconds max and that's being generous.
I'm pretty sure that I talked about the "execution time", not the "loading time".
It's not that they don't care how much time it will need for someone to see the page, but what matter is the charge implied by the page. There's more than 4 millions members, for an average around 15.000 simultaneous connections, more than 6 millions posts, without counting the private messages, and, more relevant to this page, more than 10 000 threads that can possibly be updated. Of course, processing the stats is near to atomic, and don't imply a high rise in charge for the servers. But when you works on an architecture with such size, you want to know the exact numbers, not extrapolate what they could be "without this". And you also want to be sure that a sudden slowness effectively come from the process, and not for something external.
Even with the caches, there's a high charge on the servers, and I totally understand why Sam and co want to know precisely how a page, that is probably one of the most requested one, insert itself in all this, and what level of stress it can add.