Bias in distort selection procedure?

[TheCosmicKid]

Hi,

I've noticed a bias in the Select > Distort procedure such that the distorted selection drifts towards the center of the image, with more pronounced drift near the edges and in the top-left side of the image. Weirdly, the same thing happens, in the same direction, when I invert the selection and distort that instead. Has anybody else seen this? What could be the cause of it? Is there a way to correct for it, or another function that can achieve a similar effect without drift?

Attached is an image of three consecutive distorts at threshold = 0.5, spread = 3, granularity =3, smooth = 3. Threshold doesn't seem to affect the drift. It gets stronger as any of the other three increase, but it's noticeable even at 1, 1, 1.
   

[Ofnuts]

Looking at the code, the "granularity" is implemented by scaling down the image by the granularity factor, applying spread noise and blur, and scaling the image back. I think that what happens is that the spread noise on something close to a layer edge biases the shape away from the edge because it will produce pixels mostly towards the center of the layer (the "balancing pixels" fall outside the limits and are not taken in account).