Well from samples taken from the fermenter, they're actually pretty damn close to what the STC reads. I see little point in trying to guess what the ambient needs to be at, due to the stages of fermentation creating no heat during lag phase, then ramping up in the most active phase before dying down again at the end. The ambient air temperature would need to change depending on the stage of fermentation to keep the beer temperature the same; sure, the beer temp will reach equilibrium with the air temp when nothing is happening in it, but it will rise above that 18C during active fermentation, and it will stay up there until the fermentation dies down. What's the point in that? You end up fermenting 2 or 3 or 4 degrees above what you want to ferment at.
It makes a lot more sense to me to measure the temp of the fermenting beer since that's what is being controlled, whether via a fridge or whatever other methods are available.
I use a differential of 0.3C with the probe taped to the FV under heavy foam insulation. The fridge cycles on and off during active fermentation no more than the compressor itself would if it was being used as a normal fridge, and even less often once it dies down. Which suggests to me that it is reasonably accurate as the time taken for the beer to heat up again increases as the fermentation dies down, because it isn't producing as much heat.
It works for me, and has done very well for the last 3 and a half years, so I have no plans to change from measuring what I'm trying to control, to something that basically amounts to nothing more than guesswork.
