Spent yesterday evening and this morning paralleling (and leveling for giggles and to have a starting point). I'm on a manual Rototex, so it's a bit different.
My best logic process so far - get the base near level.
Designate one platen and one head.
Test a static aluminum for true flatness on the exposure glass.
draw back the OC bolt to zero, tape a nickel to the lover part of the OC (on this press that's the camed roller-bearing arm). This gives a standardized OC and makes it so the screen would go LOWER than the surface of the platen when the nickle is removed.
use the three platen support bolts to make the surface of the platen come up and just barely meet the bottom of the screen.
[one thing I only picked up this time - THE LOCATION OF THE BOLTS RELATIVE TO THE SPACE THE PLATEN OCCUPIES - my rear bolt is past the tail by about 2" and the front 2 are almost in the middle - so you can't just think "back, front left, front right as adjusting back changes the pitch of the front too, etc.]
This may require adjusting the 2 bolts that adjust the screen arm back near the central bearing (you can adjust the tilt up, down and side to side) as well.
That's good enough, but I went a step further and made the platens level to reality as well to keep printing stresses even.
From there, add a nickel to each remaining OC base, true up each platen to the FIXED IN PLACE designated main screen arm, using ONLY the platen bolts.
Then adjust all the arms to sit flat to any given platen (at this point they should be truly interchangeable).
The only part of the logic I haven't yet solved is - is the nickle spacing a relatively good starting place to
A) be sure the screen can sit flat on the platen to register and
B) not go too far out of tilt withing the range of OC settings we mostly use (without needing to reset all 6 arms to flat).