John H and Roto - Excellent, excellent points. Upon thinking more deeply about this issue, I remembered that there actually are two relevant angles to this discussion, and that a few extra degrees of edge angle can indeed be critical.
The first angle involved is the one that the bottom of the ski makes with the snow. This is the obvious one that everyone thinks of and usually tries to maximize when they want to lock in a traverse or carve.
The second, and more critical angle can be defined in a bunch of ways, but the way that Roto just mentioned is fine: lets take it as the angle between a line drawn from the CM to the relevant ski edge and a line along the bottom of the edge/ski. Despite its extreme importance, almost no one thinks explicitly about this second angle.
Also, because few people distinguish between these two (related) angles, no common terminology has been established, specifically, which is called "THE" edge angle. Here, I'll simply call the first, angle "A", and the second, angle "B".
The reason that angle B is so important can be seen by considering a stationary ski on edge on a relatively steep hill, ie, you're stopped and balancing on one ski in the middle of a traverse.
Case I - no skidding:
For the skier not to fall over, the skier's CM must be directly above the point of contact with the snow. Now, if the skier can lift the downhill side of his ski slightly above horizontal, the little ledge that the ski has created in the snow will be slightly higher on its "air" side than on its snow side, and hence, the weight of the skier will tend to drive the ski further into this little snow ledge. He will not slip unless this little snow ledge can't support the weight of the skier and crumbles away. This corresponds to angle B being less than 90 degrees.
Case II - skidding:
If the skier lets the downhill side of his ski drop slightly below horizontal, the little ledge that the ski has created in the snow will be slightly lower on its "air" side than on its snow side, and hence, the weight of the skier will tend to drive the ski out of (and off of) this little ledge, and the ski is guaranteed to slip. This corresponds to angle B being on greater than 90 degrees.
The important thing to note is that the difference in angle A between the stable and skidding situations may be only a few degrees. If the hill is 30 degrees, angle A might be 32 deg in the stable case, and 28 deg in skidding case. Both of these numbers feel and look like a large amount of edge angle to the skier, so he may not make the effort needed to generate the extra few degrees of edging needed to lock into a traverse. (Note - the above numbers for angle A correspond to values of B of 88 and 92, degrees, respectively).
John H - With respect to your plus/minus 2 degree ankle angulation argument, my feeling is that for most people (not near their flexibility limits), its probably pretty easy to generate an 4 extra degrees of edging by more knee/hip angulation, but I agree with you that that the difference of 4 degrees between a floppy ankle and an active ankle can be extremely important. BTW, I think that plus/minus 2 deg is about the absolute maxim transverse slop you would have in modern boots.
In addition to the above effect, as was pointed out earlier, even small ankle movements within the boot almost certainly trigger or facilitate considerably more angulation further up the body.
Guys - this is a great thread. It certainly helped me clarify these issues in my own mind.
PS#1 - This discussion has caused me to wonder where exactly I (as an old fart) fit into the edging / flexibility spectrum. If I stand on hard level surface with my boots and skis on, say 8 inches apart, I can only achieve an edge angle of maybe 30 degrees (I'll have to actually measure it to be sure). This does not seem like much. How does this compare to other people? BTW, to compare apples to apples, I'm not interested in the case where you are very low, one leg is laid way over and the other is used as an outrigger. Obviously, you can get more edge angle this way.
PS#2: Here are a couple of assumptions that are behind the above discussion. I put these down here because, while important to some people, putting them in line above would interrupt the flow of the discussion.
First, for simplicity, I assumed that the metal edges of the ski have a base bevel angle of zero.
Second, for the purpose of this discussion, I've temporarily ignored carved turns or other dynamic situations. This is because the acceleration vector that these introduce must be added to gravity to determine the local, instantaneous direction of "vertical" that is most relevant to the interactions of the skis with the snow, and this would just obscure the simple discussion of angle geometry above.