IQ,
Totally agree with your test. But just as the ratio changes on the p'rod side, it also changes on the valve side because the rocker is in a new starting position when you change the prod length. The ratio is not constant on the prod side...or the valve side.
Smokey said it, & I quoted him because he was & is correct.
This [ the ratio change on both the prod & valve side ] was comprehensively demonstrated with the drawings accompanying the Crane quick lift rockers. Article appeared in Circle Track magazine Jan 06.
With a rocker arm, both the prod cup [ or ball ], and the roller tip on the valve side swing through arcs. The drawing shows A & A1 for the prod side, B & B1 for the valve side. A & B are the at rest, or starting positions; A1 & B1 are the actual positions as the rocker starts lifting the valve. A1 & B1 are longer than A & B.....& that is because the ratio changes on each side.
The text describes it:
"The relative change of A1 to A is much greater than B1 to B; therefore the rocker ratio decreases as the valve opens. This is due to the low position of the pushrod seat with respect to the low position of the prod seat with respect to the centre of rocker body rotation."
The amount of ratio change & where it occurs in the lift cycle is going to depend on the individual rocker design & geometry. This is what D. Vizard was demonstrating in his rocker tests; he tested 24 types of rocker.