Sticking the probe directly on the fermenter will give you a more accurate reading on the Temperature controller, but that does not mean a better fermentation. Look at these scenarios:
A: Temp controller is set to 2 degree differential and the probe is insulated and attached to the fermenter (or a liquid dip probe is inserted in the actual wort). You want to ferment at 65*. You set the TC to 65. The freezer will cool until the actual wort hits 65* then it stops. The cooler will then not turn back on until your wort temp goes up to 67* Once it hits 67* the compressor will kick back on and cool until the actual wort temp goes down to 65*, this will take a while and by this point the actual temp in the freezer will be a lot lower than 65*, causing the wort temp to keep dropping even after the compressor shuts off.
B: The temp probe is dangled in the air near the fermentation vessel. The same temperature parameters of above are set. The controller now keeps the fridge between 65-67 at all times, which means it kicks on more often but shuts off quicker. The wort temp itself will usually remain steady at the upper end of the range, in this case 67* The wort is kept from swinging since the ambient temp stays very consistent.
Personally I dangle the probe near the fermenter now. To ferment at 65* I would set the TC to 62* with a 3 degree differential. This keeps it at 65* the entire time. Experiment to see which way works best for you, but most importantly don't rely on the TC thermometer. You need to have a sticker on the vessel or dip in a thermometer to be sure of the actual wort temp, no matter which way you choose.