Are you sure it's a good idea to record the hottest temperature you measured in idle? Wouldn't it make more sense to get the hottest temperature at 5 minutes after startup, so that previous test runs cannot influence the result?
I did a bit of cooler testing at work, and I noticed that under certain circumstances, the CPU can retain some of the heat, making the startup temperature higher for one run than for another, even if they averaged out at the same temperature after a while.
Also, using averages over a small time period would help to reduce noise in the measurements.