Somewhere there is an error in your mathematical model or my physical one. I ran a test with my modified tension rig. The RR bridge geometry was made by affixing a bolt (black) to a hinge (grey) like this:
View attachment 3948437
I set it up so that the measurements A and B are the same as the conventional setup which looks like this:
View attachment 3948445
This ensured that the angles were the same. Both were set up with 34" scale (speaking length). The string was tensioned to a specific frequency, the same in both test cases, thus assuring the same string tension. A gauge was hooked onto the string at the point where it met the saddle/tone pin and was raised until the upforce from the gauge just exceeded the downforce enough to allow a 10 thou feeler to be slipped under the hinge at the tone pin and under the saddle height screws.
View attachment 3948471
So effectively I was measuring the vertical force applied by each setup at a string height of A+10thou.
The first (RR mockup) measured approximately 9 pounds. The second (conventional bridge) close to 18 pounds. I wasn't worried about precise measurements since I just wanted to see if they were close. And since they weren't even close I decided I didn't need to refine the setup for more accurate results.
Can you re-examine your mathematical model to see if you can account for this?