Yes, I would agree that in some cases pain compliance can lead to better results. I'm talking more along the lines of silly stuff like Hatsumi kicking a guy in the nuts 4 times after he's supposedly moved him off of mount, or doing patty-kicks to the head when he's in north-south with absolutely no effort made to control the body. To your example, I would actually argue that learning the throw by hooking the leg instead of kicking in the nuts is the better option, because not only can you train it more, but then you'll know how to throw if your groin kick fails to achieve the desired result.
Which is a joke. Reacting to possible pain reaction is a pretty terrible way to train a technique, because once again pain is unpredictable. In other words, if your trap and roll is dependent on a nipple twist, and you're in a situation where you can't reach or twist the nipple, then your trap and roll is doomed. It's even worse when your partner does a perceived reaction to the supposed painful technique and just rolls off of you like a helpless baby. That simply isn't the reality, and that's why utilizing the principles without the pain factor involved is the superior way to train the technique. If my Upa is based on sound body mechanics (proper trapping of the foot and arm, proper bridging and turning), then it has a much higher chance of being successful against a wider variety of assailants.
You can apply that same general principle to what I said above. If I've performed an Upa under duress hundreds of times in class, then I have a very high chance of performing that technique when someone is actually on top of me. Meanwhile, if I'm learning a side control escape and that escape is completely based on me jamming my finger into the eye of my assailant, I have a far less likely chance of performing that escape when someone is actually on top of me. Why? Because I've never really jammed my partner's eye, and I've never experienced the wide range of possible reactions from the eye jam.