My point is that higher resolution can salvage motion blur in some cases, and it *NEVER* hurts, at the image level.
I agree with the second point, but your first is completely incorrect. Sensor resolution NEVER alters the degree of blur imparted to the image by camera shake, lens aberration, etc. When you have 2 approximately equal blur-inducing factors at work in the same image, reducing the degree of one factor can increase overall image resolution, but that does NOT mean that you're actually reducing both blur factors.
In my examples, all the arcs were clearly resolved at all intermediate resolutions. Perhaps, in retrospect, I should have displayed them.
You have asked me what my point is, and I have answered. I still don't know what your point is. Your point seems to be that in some cases, the higher-MP doesn't help.
My point is that you are making several fraudulent claims:
1: Higher-resolution cameras are less susceptible to motion blur than lower-resolution cameras, all else equal.
2: Motion blur has less of an effect on an image as sensor resolution increases.
Both of these are demonstrably false. Let's conduct a little thought experiment, and perhaps you'll finally understant your error. Imagine a tripod-mounted camera set up next to the finish line of a race track with a 100-degree FOV. It is aimed perpendicular to the track at the finish line, and is triggered by a motion sensor so that the as a car crosses the finish line (which is centered in the FOV), the shutter fires. The shutter speed is chosen such that the car travels through 1 degree of the camera's FOV during exposure.
We start out with a 1000-pixel-wide sensor. The motion blur of the car is 10 pixels long, and if one makes an 8x10 print, the blur is 0.1 inches long on the paper.
[attachment=1452:attachment]
Now we substitute a 2000-pixel-wide sensor. The motion blur is now 20 pixels long. When we make our 8x10 print, the blur is still 0.1 inches long. While the stationary background of the image is noticeably clearer due to the increased sensor resolution, the car itself is not resolved with significantly more detail than in the print made from the 1000-pixel sensor. Why? Because the motion blur is the primary resolution-limiting factor, not the pixel count of the sensor.
[attachment=1453:attachment]
If you compare my two sample images, you'll note that the car is virtually identical between them. The length of the motion blur is identical, the difficulty of reading the lettering on the car is identical; overall the differences are very subtle. The only real improvement to be found is in the horizontal lines of the car (top/bottom of the window, grille, etc.), which are not affected by the motion blur. Other than that, the additional sensor resolution has not helped resolve the car any better, because the motion blur is a much more significant factor than the pixel-size blur in both images.
The notion that more and smaller pixels requires more steady technique is nonsense; it is only true if you're going to make a MP-equivalent crop of the low-MP version in the high-MP camera, and expect them to compete at the same viewing size.
It's far from nonsense, it's inescapable physics. The whole point of increasing sensor resolution is to capture more overall image detail. When camera shake, motion blur, focus errors, or lens aberrations degrade resolution to a greater degree than pixel-size blur, the advantage of adding additional sensor pixels is compromised or for all practical intents eliminated. The smaller your sensor pixels are, the easier it is for other blurring factors to overwhelm pixel-size blur and degrade image quality to something far less than it could otherwise be.
That is why people who compared the 1Ds and 1Ds-MkII generally found less of an image quality advantage by upgrading to the 1Ds-MkII than might be expected solely from the difference in pixel count and the generational improvement of the 1Ds-II's sensor. The 11MP 1Ds is already quite demanding on most available lenses, and a lens that struggles to satisfy the 1Ds has an even more difficult time meeting the demands of the 1Ds-MkII. When aberrations, whatever their cause, are > 1 pixel in size, there is very little benefit to throwing more pixels at the problem. The more significant the aberrations are, the less significant the benefit derived from the extra pixels will be. And the smaller the pixels are, the easier it is for the aberrations to negate the benefits of additional pixels.
If your assertions were correct, then handheld MFDBs would be in common use for shooting action sports. Perhaps you should consider why smaller, lower-resolution formats are most commonly used for such tasks.