You can use the onboard sensors as a delta/starting point but there's no way they're actually accurate. They may be precise and predictable but far from accurate (real temperature).
As a casual user, the only way to test coolers in your system is to be diligent. Compare idle and load temps of your coolers with the same ambient temperature. Ensure proper seating and thermal paste application with both coolers. You cannot compare temperature numbers from someone else to your own, as there's no control. Basic experimental procedure.
Any hardware reviewer worth his/her salt uses hardware monitoring. I've seen a few places drill an actual sensor into the casing of the cpu.
There's also no way an air cooler gets you close to within 2 degrees of ambient (or -1?? WTH??) I can't believe you used that as a point of performance.