A couple of students were being introduced to a new technique: feinting from change-of-guard. M explained it, and I let them play with it for a while, before going over to them and explaining that it all relied on anticipation. This confused them. So I explained.
Now, the guards in question are outside and inside guards. If you're right-handed, and you imagine you've got a large box in front of you at chest height, then outside guard is where the hilt is at the lower back right corner, and the point is at the upper left front corner. Inside guard swaps so that the hilt is at lower left back, and the point at upper right front. When you engage blades in this position, your opponent mirrors your stance (rotationally). When you change between the two guards, - tipping point back to vertical in the front-to-back plane, moving hand and point, and then dipping point forward again - your opponent mirrors again.
The technique is thus: at the point where your opponent mirrors your change, you don't actually complete the change, and instead whack him in the side of the head ('scuse the technical jargon).
The thing is, although the box analogy gives you a rough idea of the positions, the timing of the swords' relative motions and the reasons for them are surprisingly complex. I explained all of this to the two students, and then went through the process in detail. They'd heard this before, but hadn't been at the level where it made sense, so it hadn't sunk in. Now they were, so it did.
Fiat lux. Ding!
Now they realised that when they'd been "changing guard", they'd just been (literally) going through the motions, and suddenly, it all made a lot more sense, and was a lot more difficult than they'd thought. Never mind the new technique, they decided; they wanted to practise this fundamental a lot more, until they could do that. They spent the next twenty minutes spontaneously developing exercises to develop and practise their new understanding.
And then they went back to the new technique, with a new-found understanding that it relies entirely on the opponent not really paying attention to what they're doing.
Now, all of this was pretty rewarding in itself (the ding was damn close to being audible...), but I'm particularly chuffed because I remember discovering that this guard-thing was far more complicated than I'd thought, back then, but forgetting why; so I've spent a looong time thinking through the mechanics and justifications purely so that I could explain it clearly to a newer student.
It was worth it. Very gratifying.