One way to do it is to experiment with models. At least that will give you a good idea of
what is going on with the impedances.
You can model the antenna using EZNEC, 4NEC2, or any of a number of other modeling programs. If you
aren't already familiar with them, you can also use W9CF's online yagi antenna modeling applet with a
single element: set the element length to 102', use plenty of segments, and get the feedpoint impedance
on each band of interest:http://fermi.la.asu.edu/w9cf/yagipub/index.html
Then you can enter the feedpoint impedance into VK1OD's transmission line loss calculator and see
what the impedance will be at the balun:http://www.vk1od.net/calc/tl/tllc.php
This will also let you see the effect of varying the line length and its effect on the SWR on the coax.
You can then take it a step further and feed the calculated impedance at the balun back into the
VK1OD calculator for the 50 ohm portion to see the losses and the actual impedance at the tuner.
From there, W9CF has a handy applet that calculates tuner losses:http://fermi.la.asu.edu/w9cf/tuner/tuner.html
Using these tools will allow you to see the total losses in the tuner, coax, and open wire line on
each band. It may be a lot of work to try all the different combinations for each band, but still
faster than cutting lots of short pieces of feedline.