I think this is incorrect, or at least oversimplified. With even a little bit of pressure the strop loses its parabolic shape and describes more of a very shallow "V", with the strop bending to varying degrees at both the spine and the edge. The bending at the spine doesn't affect the edge at all - if all the pressure is focused on the spine and the blade just wafts behind, then all the bend in the strop is at the spine and the edge will strop normally on the straight trailing edge of the "V" - you're just using the spine to finish tensioning the strop. Similarly, even with a shallow sag you can dull a blade if you apply a torque to the blade so all the bending in the strop occurs at the edge of the blade. Of course it's easier to get everything right if you minimize the variables, which is where the "keep a taught strop and use very little pressure" rule comes from - it's easy to keep the strop taught, and easy to keep the pressure light. Or at least easier than dealing with the changing angle, pressure, and torque as you stroke across a slack strop, stropping against the straight trailing edge of a changing angle as you sweep across. It can be done - we've all seen Lynn shaving off a slack strop in his youtube video

But the fact that something *can* be done a certain way is not a reason it *should* be done that way.