...pumps. In your rig there's a large angle between the link and the axis of the "tube" and direction of the 500# load. This angle is (much) worse in the case of the short link than the long link, which in the former case is placing a side load on the "piston" that's (much) larger than the load itself. Consequently a large part of the resistance you felt at the lever in the short-link case was probably caused by friction between piston and tube.
But in an actual pumper, by contrast, pressure and force don't become significant until late in the stroke when the lever is far past vertical and the link, no matter what the B/A ratio, is nearly parallel to the tube, making the contribution to total effort that comes from side loading and friction relatively minor even for short links.
To cut to the chase, and unless I'm still failing to visualize it (very possible), I guess I'm saying your whole test was B.S.