That article is accurate, but does not pull the pieces together. Let me fix that.

We all know that hot air rises, and cold air sinks. Therefore if you heat one place and cool another, air will rise in the one, move to the other, and then sink. The Earth is heated up near the equator, and cooled near the poles, so we show that pattern. It actually repeats several times, and the result are the Hadley cells. See http://en.wikipedia.org/wiki/Hadley_cell for confirmation of this.

However in the Earth there is a complication. We are rotating. When air rises, then moves towards the pole, it maintains its angular momentum, and now goes a lot faster than the ground. Thus the air that is about to come back down is going very fast. We call this the jet stream. (There are actually a few jet streams at different latitudes.) When it comes down, the ground slows it down, but it results in a prevailing wind. Over the oceans those winds are called the trade winds because sailing ships used to follow routes that used them. The one over the US goes from west to east, and that is why many major storms tend to do that.

Next piece of information. The nice symmetrical structure I've drawn so far is unstable. It is prone to develop kinks and bends, which then grow. Around the world the overall average latitude is reasonably well fixed, just not the exact latitude at a specific spot. The result is that the jet stream tends to develop fairly large standing waves which are fairly long-lasting but do move around.

Now go get a globe and look on it. North America turns out to be roughly half a wavelength for one of these standing waves. So if the jet stream goes north on the Pacific side, it will go south by the Atlantic. This is our current pattern and it results in a mild winter on one side of the country and a severe one on the other. If the wave is flipped in another year, the pattern reverses. Or you can have a phase shift and both coasts can be average and the interior can be either warm or cold.

This has been known for decades and has nothing directly to do with global warming. (I learned it all from a fluid mechanics course that I took in 1992.)

"Directly" is practically meaningless when discussing cause and effect in a system as complex as the global climate. But there is ample evidence that arctic warming, for example, can profoundly alter circulation patterns in the upper atmosphere, which lead to extreme weather events in more temperate regions (http://skepticalscience.com/jetstream-guide.html). Personally, it's a lot easier for me to buy that the climate has altered than that I'm witnessing a true once-in-25000-years weather event.

The jet stream, being chaotic, can be affected by anything.

That said, the locations of the Hadley cells changes depending on how much heat is stored in the system, and adjustments to this is part of the whole El Niño, La Niña cycle. As you increase the rate of energy storage, this results in various chaotic changes, that can increase extreme events.

But the real problem with the rarity of the event is that you arrive at that figure by a probability model assuming independent storms, probably with a Poisson model for frequency. However storms are not actually independent events. If you have a stable pattern that results in many storms following the same track to Boston, the odds of an extreme event are way higher than the naive probability model would have suggested.

Agreed, but the more problematic assumption is that the Poisson rate function is time homogeneous is what I'm saying.