You cannot check only once and then conclude anything from that. Tor
network is really a place where you should do a statistical research
as is almost as far from deterministic as it can be.
During last year of my playing with Tor network I did plenty of test in different days, random circuits, fixed-order circuits and so on, of course. But Im just surprised by speed in ONE random test with wget, because I know, that is far away of typical behaviour. It is all. And the "conspiracy" was bad joke, of course ;-).