Some engines with certain cylinder heads (like dart) have mediocre to bad valve train geometry to 8 valves at best. This can be corrected with custom rocker arms half sets from like crower. I was building a 565 with profiler heads, if i remember right the intake geometry was terrible. Was what I call a "10 hr rebuild engine" , ie, was going in something that would maybe see 10 hrs run time before its next rebuild being abused in its application. Had a cam that would result in a about .800 lift. Customer would NOT buy shaft rockers and we couldn't get the custom crowers by our deadline. So after rolling it thru its cycles a million times times looking at its bad geometry there was about a .150 range of pushrod that gave it its "best" bad geometry. I put a dial indicator on top of the retainer and repeated rolling it thru in .025 increments. surprisingly I found the net lift varied from about .745/.750 to .785 or so, I went with the length that gave me the most net lift FWIW, Smitty

The 90 degree comment is that valve train doesn't run 90 degrees anywhere. It's always off. Adjustable pushrods and sweep will get you where you need to be and then order lengths. It will change running anyhow I find.

I kind of wondered if maybe a backset rocker would work properly with my AFR heads (they’re a v2 head casting so their rocker stud flaw was fixed). I could never get the mid lift method to work - the witness mark would end up too far on the exhaust side and the pushrods too long. I’d also get gross lift numbers that were inconsistent. When everything went together and I ran them, the valvetrain was really noisy. After trying to adjust them a couple of times, I had no change - at first I thought I had lifter problems caused by excessive spring pressures (Crane tall body), but when I dug into it and searched for similar experiences with these heads and the mid lift measuring, I remeasured everything the old fashioned way - end up shortening everything about .050 and boom! - quiet valvetrain (for a BBC).