Based on some (mis)analysis too obscure for mortal men and women to follow...

It's not so obscure. Goddard simply did not understand how a map projection could affect his numbers.

A map projection is a way of representing the 3-D Earth on a 2-D piece of paper (or a screen). The map projection a lot of people grew up with is Mercator. Check it out:

Mercator projection

The Antarctic is HUGE -- bigger than all the other land masses. That's because the standard Mercator projection distorts the area near the poles.

Now, just for fun, here's a different map projection:

Lambert Azimuthal Equal Area projection

This one shows the size of the Antarctic in correct relationship to the size other land masses. But the shapes of the continents are distorted.

In fact, all map projections have some form of distortion. It's unavoidable.

But Goddard simply counted the pixels on one map projection, compared it to different map projection, and ignored the fact that they were different.

That's why the NSIDC says,

Such an approach is simply not valid.

The proper way to calculate a comparison of ice coverage is by actually weighting the pixels by their [area] based on the map projection, which is exactly what NSIDC does.

By the way, who is Steven Goddard and what are his credentials, training or background? It seems to be a mystery. Some bloggers speculate he is a pseudonym.

Ped Shed Blog

The ice extent plotted by NSIDC has taken a nose-dive in the latest plot. This is kind of wierd, the decline should be leveling off at this time of year, but has accelerated in the past week.