If you are handy with calculus you can figure out the location that gives you the lowest combination of load and travel. The Mid Lift setup gives the minimum travel across the valve which is fine for low load springs. But if you have high load springs, the load at full lift is much greater than it is at low lift so you might want to reduce the amount of sideways travel near full load.
Mid Lift gives you max travel at full load which might not be good for things. It takes some advanced math to figure out but you can minimize the product of load and travel by moving the shaft down. You do end up with more travel across the valve, but that extra sideways travel occurs at the beginning of the lift cycle when the rocker arm is lightly loaded.
This is Jesel's philosophy for their rocker geometry and why I won't use their rockers. The problem with this is that it changes the acceleration rate of the valve and if I want to do that, I'll do it with the cam. I don't see the load differences being worth the trade off and I want the rocker to be as neutral as possible. I've had this conversation with just about every engine builder I come in contact with since I'm so intrigued by this and I can say even at NHRA/IHRA Pro Stock level, there are those who prefer mid lift, and those who prefer what you describe. Bottom line here is either are very close to one and other and unless you have the
exact rocker height for your combination, and the pushrod side has the same perfect geometry, you're going to end up with a variation. Using mid lift is easy and basic and will get you a whole lot closer than "centering the pattern on the valve tip". There are also load dynamics that play into all this that take a little deeper understanding of what's going on but for the most part I prefer to keep it simple. At full lift, there may be zero load on the rocker and under (lifter) acceleration, there may be way more than spring load on the rocker. An ex rocker is trying to open the valve against cylinder pressure. Just other things to consider.