I think the reason there's some hesitation, is that retrofocus lenses behave a little strangely, and you have to start thinking about entrance and exit pupils, if you calculate it in that way.
My suggestion: take focal length out of the equation. I think this should work independent of the issue of the design of the lens, and it's the formula I use. It works for me, but then again, I don't think I've ever had to reverse a retrofocus lens, but a quick test should tell you if it's right.
Change in exposure=(1+magnification)^2
where a change in exposure of 2x will be 1 stop, 4x is 2 stops, 8x is 3 stops, etc.
An attraction of this method, is that it works with any format. I have a magnification/focal length table taped to the back of each of my LF cameras and in my photo notebook for other cameras, so I can quickly estimate the magnification by comparing the width of the scene to the width of the film frame and convert it to an exposure factor in f:stops, or I can measure it more precisely by putting a ruler in the frame and seeing how much of it I can see in the finder or on the groundglass.