I've noticed that, at least on street motorcycles, the tread elements are oriented length-wise with maximal combined (cornering plus braking front, cornering plus drive rear) loading. That means the front's "arrow" pattern points rearward.
On most off-road bicycle tires, tread elements are oriented cross-wise with braking-plus-cornering load, both front and rear.
Road bike tires seem to orient their tread elements either way, depending on brand. Length-wise orientation solidifies the grip to smooth, clean pavement by preventing tread distortion. Cross-wise interrupts the effects of small amounts of gravel, that is to say that the sand/gravel's effect on traction is a bit less continuous as the rolling gravel bits must pass across the tread's sipes. Thicker and/or softer rubber, and lower pressures, similarly lessen the ability of gravel particles to lift most of the tread's surface clear of the road, thus improving traction over dirty surfaces.
The net effect of the very shallow tread as used on racing tires would seem to be making the tread more compliant to rough or dirty surfaces, simulating softer tread in some ways. Note that much of the tread on road tires is clearly so far of center as to never actually touch the road, except perhaps when crossing pavement seams that are oriented fore-aft, as is often the case near the road's edges. It also appears that the centermost tread still contacts the road during hard cornering, unless pressures are run ridiculously high. Cornering loads pull the center of the tread significantly toward the direction of lean.
Hydroplaning simply does not occur with bicycle tires, but any tread which speeds up the "drying time" would seem beneficial after, say, crossing a lone puddle. Note that water flies off the center of the tread, though, regardless of any fine tread's pattern, but on continuously-wet road surfaces, no drying occurs.
Note lastly that my off-topic (23mm) Vittorias have a directional tread, but are labeled (loudly) on both sides!
David Snyder
Auburn, CA, usa