Missed on this combo?

Just ran some rough numbers... I guessed at the tip-to-tip rocker length as being about 2-3/4" so if that is off, then the number are off in proportion. And note that this change of lift at the rocker tips due to the rotation occurs at both ends; the vertical motion of the pushrod end at the rocker gives a bit more rotation than would be expected, and so somewhat compensates the loss of at the valve end. (For a 1:1 rocker, it would all cancel out.)

With my rough numbers, it looks like you are losing about .003" to .004" more just from and 'ideal' design at these lifts, but the ball position makes it worse as noted. So now you have lift losses roughly in the range of:
- .010" from the lash (Thanks for the correction! duh...)
- .012" from push rod angle
- > .004" from general design of the rocker

So roughly half of your loss of lift is just in these things assuming they gave you a valve lift at a given rocker ratio. I have not looked at how your cam is spec'd; if they spec just lobe lift, then you use the rocker ratio times the tappet lift minus these corrections to get to the real valve lift. Now if they advertise a specific valve lift, then it ought to be there. (I suspect that most cam mfr's just multiply the lobe loft by the rocker ratio and give that as advertised lift, but I don't know that as a fact.)

And for sure, the .030-.035 roller motion on the tip is right for the 'best you can do' correction. Again, this is just naturally from the design of the rocker.