don't agree, just opinion:

i think the rocker should give an idea of what the curvature of the bottom is. if you use 2 horizontal measurments you have no idea of what the bottom looks like. take your boat for example and from the side proportionally increase the depth only over the whole yak and reduce its lenght. for the same load you could end up with the same figures with radically different side profiles of the bottom.

the depth has to be part of the measurement not just the proportionally of where it lands.

for me, any length is semi-ok. the 50% waterline or where 50% of the depth is above or below - don't care, but then that DEPTH measurment must be included.

so in your example 16' - 1.5' -1' = 13.5' = 162"
and depth is 2.5"

so just divide to get proportion and curvature of 2.5"/162" (this is ok, but then)
take percent for ease of understanding:

2.5"/16" x 100 = 1.5% rocker

now you can draw this profile out for any legth size vessel to compare.

like before i personally like the 50% of the waterline length as then you know you are away significantly from any stern or bow idiosycracies. a good example of this indiosycratic behaviour is the dropped skeg protruberance on many skin on frames. the 50% depth then gets pushed back, but the question is whether or not you/we wish this to affect the rocker measurement(good can of worms)