It depends on what you are trying to control.
If you have your STC connected to your fridge, then you are controlling the air temperature in the fridge . . .therefore it makes sense to have the sensor suspended in the air in the fridge. Due to the way fridges work and the fact that air is an excellent insulator of heat then you will get swings and variation . . .That's why commercial fridges have fans in to ensure that the temperature is even throughout the fridge
. . . Simple solution for us put a fan in the fridge.
Now wort behaves differently to water, and if you are controlling the temperature of the wort, then you need to have the probe immersed in the wort. Please note that sticking a fermenter in a temperature controlled fridge is
NOT controlling the temperature of the wort in the FV. As you have (hopefully) a stable temperature environment then you should have a stable fermentation temperature (the mass of liquid helping to cushion against the minor swings inside your temperature controlled environment. However, with a decent healthy fermentation do not be surprised to see your wort temperature up to 5C higher than your ambient temp . . . Ok Not a problem says you I'll tape the probe to the FV and insulate it from the air. This will give you a better measure and control of the wort temp, and indeed it does . . .but you have now decoupled what you are controlling from what you are measuring.
Thankfully because of the laws of physics (and ye canna change the laws of physics) the devices that we a re using to supply heat and cold are relatively low power, and the insulating capabilities of the air in the fermentation chamber, and the heat capacity of the liquid to resist changes in temp then the swings are damped to quite a great extent.
My personal choice is to have a temperature controlled environment (with probe in the air), and Direct control of the wort using a coiling coil and maxi chiller with the probe in the wort . . . with my 90L batches I get swings of around 0.2-0.3C . . never more than 0.5C.
Smaller 23L batches sit in a water bath with a coiling coil driven by a maxi chiller,and a 100W aquarium heater. as the mass of water in the cooling bath (~23L), is the same as the wort and the specific heat capacities of the two liquids are very similar, the probe sits in the water bath. I use a small pump to recirculate the water bath to equalise any temperature stratification, and to ensure that any heat generated by fermentation is taken away from the side of the FV.
Control is to within +/- 0.5 of the Set point.
The good thing is that I can set the STC to have a 60Sec delay (5F IIRC) between switching heat / cool which also helps to prevent over swings in both environments. . . Those of you using F4 to compensate for the difference between wort temp and fridge temp (what is being controlled) should be aware that you may potentially introduce swings (which may be significant), as this setting is used to compensate for the deviations measured between different probes.