1) effective diameter = shortest dimension
2) effective diameter = diameter based on total area

I tend to think that 'effective radius (shortest distance
between front and rear of driver)' may be more appropriate.

But people say that because an OB is a 'velocity' rather than a
'pressure' system that the circumferance rather than the (true)
diameter sets the LF cutoff ????????????????????????????????

But if the baffle isn't circular with the driver in the center, transition is staggered and more gradual. Defining an effective diameter for any baffle shape other than circular is a simplification. If you consider the shortest dimension then you'll probably have a little more bass energy than you expect and if you average area, you'll probably have a little less.

As I understand it (not that that means much

The 2 schools of thought......Say we have a 1' diameter baffle:

1) the circumferance must = 1wl so.....
cir = pi * dia = pi
Fc = 1130 / pi = 360Hz

2) the distance front to rear of driver must = wl/2 so
Fc = 1130 / (2*dia) = 565Hz

#1 is certainly more attractive and seems to work, but what about the 'rear wave cancellation' of #2. ???????????

That's what I need !

So say a 3' x 5' baffle on the floor.....
a = area = 15 sq.ft.
r = radius = (15/pi)^0.5 = ~2.19'
d = 2 x r = ~4.37'
Fc = 565/4.37 = ~130Hz
Since it's on the floor Fc = .707*130=~92Hz
-3db = .707*92=~65Hz
Sound reasonable ?

It might be interesting to measure several baffle shapes and develop a mathematical model to describe the response. It would be easy to do with cardboard and could help you make a spreadsheet to predict response.

Yea ..... But .....
Here's my (tentative) plan:
Use a widerange driver to get down to say 100Hz.
So a 100Hz driver with a Q of 1 will give a flat
response (Q=~.7) on a 100Hz baffle.
Then add a 'helper woofer' to extend down an octave.
(rolled off on the high end with a 12db/octave passive
network)
(a driver with a Q of 2 @ 50Hz on a 100Hz baffle is
flat down to 50Hz)
Thus the need for a fairly good definition of baffle
cutoff since the baffle is used as a crossover element.

You might makes some test baffles with cardboard of different shapes and configurations and do some measurements to test everything out. I'd experiment with smaller baffle sizes to shift the frequency up so measurements would be easier. Maybe try 9 inches to a foot, using a small midrange driver. The higher in frequency you go, the smaller the distances involved and the easier it is to gate out reflections from walls and stuff. Once you have defined properties, you could scale up to the size you want and go with a larger driver for your 100Hz model.