It does make sense, Matt, but maybe a little drawing will help in this instance …
Suppose you add a stroke to a rectangle shape, centered around the rectangle path, just as you did in your drawing.
Furthermore, imagine a circle drawn around a corner point A of that rectangle, such that the diameter of that circle equals the width of the stroke applied to the rectangle. Then the radius of that circle will stand in a ratio of 1 : 1 to one half of the stroke width. See the purple lines in my illustration.
Now, when you set the mitre value of your stroke to 1, your stroke will be cropped by a tangent to the circle that is perpendicular to the diagonal from the corner point A to the corresponding corner point of the stroke. The purple non-dashed line of my illustration is part of that diagonal, and the radius of the circle is the geometrical equivalent of the mitre value in our case.
And it is basically the same, when you apply a stroke to a non-rectangular polygon. See my second illustration.
So when you increase the mitre value, you are basically increasing the radius of the circle. And when the ratio of the radius to one half of the stroke width becomes greater than 1.41 : 1 (which is the approximate ratio of the diagonal of a square to the side of that square), your stroke won’t be cropped anymore. Now, since the rectangles in your drawing are so small, you will have to go a bit further than the ratio of 1.41 : 1 in order to compensate for rounding errors, I guess …
Hope that is not only technically correct, but also makes sense …
Alex
