Here's the scenarios:
a) Probe in wort.(or taped to side of fermenter) Temp set to 17 with 2 degrees variation (ie compressor cuts in when temp is 19 and out when 15). This means your wort will slowly vary between 20 and about 14, because the coolness of the fridge will keep chilling the wort even when the compressor has cut out, and when it cuts in, the temp will keep rising at least a degree before it starts dropping. Moreso with early fermentation.
b ) Probe in air space Temp set to 16 with 2 degrees variation. The wort will reach about 19 before the probe reads 18 and kicks in the compressor. The compressor will then chill the air space to 14 before cutting out. The thermal mass of the wort will prevent it dropping this cool, maybe 16 before it stabilises with the air temperature then the exothermic behaviour of the fermentation will start the cycle again.
I'm not religious about this, tho. If someone could convince me of the merits of attaching the probe to the wort (or using an internal sensor), I'd change what I do, but I think my theory is sound.
SCENARIO A
Probe Inserted in Wort.
Small fan to circulate air inside the fridge/freezer unit.
Set temp 17*C
Unit will turn the compressor on when the wort hits 19*C
Unit will turn the compressor off when the wort hits 16*C
The air temperature of the fridge responds quickly if a fan is inside the unit, so when it kicks in the cooling effect will be reasonably immediate. With a fan running in the unit you are also adding a little bit of heat to the equation but probably not enough to make any difference. However you are correct you can overshot the target temperature especially early in fermentation, eg the compressor turns off at 16*C and the cool air in the fridge keeps chilling, the over shoot will depend on a number of factors, is any heat being generated from the wort, thermal conductivity of your fermenter, volume of wort vs cooling capacity of the air inside the fridge. The over shoot, or under shoot of temperature is the biggest concern, especially if you are dropping the temperature of wort that started at say 24*C, best monitored for the first couple hours.
SCENARIO B
Probe Inserted in airspace.
Small fan to circulate air inside the fridge/freezer unit.
Set temp 17*C
Unit will turn the compressor on when the air temperature hits 19*C
Unit will turn the compressor off when the air temperature hits 16*C
The air temperature responds rapidly with a fan inside your unit (you can stand their and watch it drop), however just because the air temperature arrives at the target temperature does not mean your wort has. If you pitched a little warm, your fermenter could still be at the original temps, as I stated it does not take long for the air temperature to adjust. If the fermentation is producing heat, you will need to wait until the heat produced is sufficient to warm the air, to enable the compressor to turn on. The system will settle and move towards equilibrium, eventually. However it is early fermentation that plays a key role in fermentation flavour by-products.
In short Scenario A is better for early fermentation IMHO (people are free to sick there probes wherever they like). And Scenario B is better for long term lagering. After crash chilling I simply insert the probe for early fermentation and pull it up a bit and leave it in the airspace inside the fermenter (probably middle grounds between scenario a and B) for lagering.
In the early stages I want to know and measure the fermentation of my wort, not the air temps of my fridge which after awhile and all the action is over will reach equilibrium the wort.