I agree with the thermowell but completely disagree with the glass of water solution. A glass of water is the best way to allow for the specific heat retention of water to properly reflect the temp of the wort and keep the compressor from cycling on and off.
Sorry, but you're mistaken. Putting the probe in a jar of water is the worst possible solution. The thermowell is the ideal solution.
Here's why: Think about what happens during fermentation. You've got your fermenter sitting in the fridge, and the beer is at exactly the right temperature. The yeast start working, producing heat. The beer warms up a little bit. If you had the probe inside a thermowell, then the temperature controller would know immediately, and could turn on the fridge to keep the temperature where it's supposed to be.
But your probe isn't in a thermowell. It's in a jar of water, which hasn't changed temperature (because there are no yeast in it producing heat). So the beer warms up. This warms up the surrounding air in the fermentation chamber. If your temperature probe were dangling loose, this is the point at which it would notice an increase in temperature, and could turn on the fridge to cool things back down.
But your probe isn't dangling loose. It's still in that jar of water. Now that the surrounding air has started warming up, the water in the jar finally also starts warming up. Meanwhile, the temperature of your beer is spiking, because the yeast are getting more and more excited, and the temperature controller still hasn't yet turned on the fridge to cool things down.
Finally, the temperature of the water in the jar increases by enough to trigger the temperature controller to activate the fridge. But by now, who knows how warm your beer has gotten in the meantime.
Now, a similar problem occurs while cooling down. The temperature controller notices that your jar of water has warmed up a little, so it turns on the fridge. The fridge cools the air down, which slowly cools down the water in your jar. It cools the beer down a little bit too, but not as much, since it a) has more thermal mass, and b) has active yeast in it producing heat. Before too long, the water in the jar has cooled back down to the target temperature, so the temperature controller kills the power to the fridge. But the beer is still too warm. The air inside the fridge will be considerably cooler than the water in the jar, so the water in the jar might continue to cool for a while, before bottoming out and starting to warm up again. During this whole time, the beer is still above the target temperature, and getting hotter again.
As you can see, this is a terrible solution. Best idea, probe in thermowell in beer, second-best idea, tape probe to the outside of the fermenter and cover it with some sort of insulator (foam, bubble wrap, whatever), third best is to just let the probe dangle freely in the air. But immersing it in a separate liquid is the worst possible solution.
The compressor cycling too much is not an issue, and even if it were, the way to control that is by changing the cycling delay setting on your STC-1000 (i.e., mine is set to 10 minutes, so if the freezer was "on" within the last 10 minutes, the STC-1000 will not turn the cooling circuit "on" again until 10 minutes have elapsed, even if the temperature has climbed above the setting).
