Just going from memory here, but...
1. Before each flip, there is a substantial weakening of the strength of the field. We've seen a few little dips lately, but nothing taking us anywhere near the ranges associated with a full flip.
2. I've become convinced that there is a significant problem with the way "due" is interpreted for things like this.
Recall our discussion in the Interesting Laws thread and how this touched on human heights which cluster closely around certain values, against river lengths which are distributed randomly, but seem to have more beginning with low first digits.
People with brains less awesome than Clodfobble's might like to read this
http://www.dichotomistic.com/hierarchies_fractals.html for preparation, especially the bits about power laws and scale free networks.
Human heights are not scale free. Imagine getting hundreds of people to lie down head to toe in a straight line. Walk along, and see how regularly you get to a head. About every 1.6 to 1.9 metres, you'll get one. In extreme cases you might range from 0.4 up to maybe 2.3. So once you leave one head the next one is "due" in about 1.7 metres. After 1.8 we're expecting it soon, after 1.9 it should be here really soon, and so on.
Earthquakes, magnetic flips, the lengths of interglacial periods and many other things follow power laws - they range randomly over many orders of magnitude in size and/or interval. You might get ten magnetic flips in a million years, then none for the next nine million. They'd happen at an "average" of one per million years, but after, say, five million years, they'd seem terribly overdue. They're not, really. It's just that average intervals do not work well with phenomena like these.
The current interglacial is a good example. There have been (I think) about 35 interglacial periods since the current ice-age began about 3 or 4 million years ago. The *average* duration of interglacials is about 10 or 11 thousand years, but they are NOT clustered around a bell curve like human heights. Some were only a few hundred years, some were 40 thousand or more. Trying to predict the end of the current interglacial by looking at simple averages like this, well, we're "overdue" for it to end already.
Same with earthquakes. There are LOTS of fault lines that are allegedly "overdue". Same with volcanoes. This is what originally made me suspect there was something wrong with the idea of "due".
Oh and also, if the magnetic field does flip, we're screwed and the only thing we could do is stock up on tinned food and a hand powered can opener!