Valve train geometry

If pattern was off side to side too much I'd move the shaft/ stud never tried (well once) centering the pattern
Makes sense to me. With a roller the force into the tip is straight down into the valve stem tip as long as the tip end is 90 degrees from the stem axis. Moving the roller off of center one side or the other does not change that force direction. If you get too far off center the roller and tip pressures increase too much but as long as they are near center, they stay pretty constant.

It might be possible to get some sort of rotational motion imparted into the valve (as viewed from the side) when off center with very high lift speeds, that the guides have to control, but I don't have any way to quantify that. But even then I bet spring vibrations and movements nearing resonance far exceed anything like that.

Narrowing the pattern may reduce the low roller side force on the stems due to lower roller-on-axle rotating speeds.