Spoke orientation is a difficult thing to conceive. The spokes themselves are tensioned sufficiently to take force in compression. The force components are radial, lateral, and tangential.
The radial force component is the wheels ability to carry the load. It’s transmitted from the frame and fork, through the axle down through the tensioned spokes to the ground. There is some force in tension, but not very much. (this is a brain bender)
The lateral forces are side to side forces that come from turns, springing, leaning the bike over somewhat, etc.
The tangential forces are from the driven components, the cogs, as well as rim brakes and disk brakes.
Tangential forces applied to the hub, either braking or driving, are transmitted through the spokes both via compression and tension. The greater the number in the cross pattern, the more leverage the spokes have against these tangential forces. This is why it’s recommended to use a 3x pattern on driven and disc wheels.
The problem with the cross pattern is that spokes in that orientation lose leverage against the lateral forces acting on a wheel. The most optimal orientation of spokes against lateral forces is a radial pattern with the heads in. This is why rim braked front wheels are typically laced with a radial pattern.
There’s some argument that heads out is more aerodynamic, but I believe that’s poppycock. It’s all a big mess of turbulent air in any case.
I would say these are the fundamental requirements driving the design of a lacing pattern. From everything that I’ve read on this topic over the decades, the only concern with heads in vs. heads for a crossed lacing pattern is whether or not the base of the spoke becomes deformed, which will weaken it somewhat.
That said, I’d be interested in reading PVD’s take on the subject. He tends to be a pretty sharp cookie.