I've read many accounts of watches running alternately fast or slow, by varying degrees, depending primarily on how they are stored - given that a watch on the wrist will average many varied positions. I'm wondering if a given common movement (Valjoux 7750 in my case, but I'd be interested in other examples as well) will show the same dependence on orientation in any and all watches that contain it, or if the dependence is more random.
I next wonder if this will depend on the "grade" of the particular movement, since I've read that the higher grades are regulated in more and more positions, to tighter specifications. This seems to suggest that a low grade of a particular movement may exhibit strong positional dependence, if it is placed in an orientation in which it was not regulated, whereas the high grade may show little or none (and be more accuarate while it's at it, which I guess is what we would expect).
I'm currently collecting data, and intend to eventually do the experiment, but I was wondering what peoples' knowlege and experience was. I hope to impact and improve accuracy with my nightly storage, and if the Valjoux 7750 is known to behave in a certain way, that information should make things much easier.