Just to add my own theory to the super-cooling debate... Imagine if you had a high waterfall that the wind blew through, or a waterfall that cascaded down across a large pile of rocks before entering the pond. What if you had the waterfall at the top of a long shallow stream? Each of these situations could cause the water to lose heat much faster than normal. Remember that heat loss is directly related to the exposed surface area of the water -- water poured over the face of a rock in a thin sheet has a much greater exposed surface area than water pouring out of a pipe. The layout of your pond has a huge effect on how fast the water might lose heat. A deep pond with straight walls will retain heat better than a shallow pond that slopes up towards the edges. And one last factor to consider is the flow of water... You have a pocket of warmer water at the bottom of your pond, but if the water from your pumps is stirring up that water, then your fish will suffer.
In my pond, the waterfall only disturbs the top ~6" of water, so the fish laying on the bottom feel very little effect from the flow. All of my pipes are buried at least a foot, and the waterfall flows as a solid stream of water over the top ledge and straight back into the pond. There isn't a whole lot of room for heat loss, so I generally only have to run heaters for 1-2 months. I wasn't thinking of heat loss when I designed my pond, but I had to keep everything compact and I lucked out that it works well through the Winter.
@addy - I can't imagine a temperature swing of 30F would be enough to hurt the fish by itself. There must have been something else that contributed? Did that guy have a raised pond or something else that contributed to heat loss? I say this because in my area, it is very common to see temperature swings of more than 50F in just a few hours as different storms blow in. Just a couple weeks ago we had the temperature swing from 8F to 53F in 3 hours, but it barely affected my water temperature.