Crow, you have listed just about everything.
A longer valve has more area that needs to be lubricated, so less would be faster.
The springs have to move the mass, so reducing mass makes them faster, perhaps at the expense of that "smooth" feel you get when pushing them down.
The actual finish of the valve surface determines how well the oil "sticks". Most new valves have been lapped in and are not shiny, rather have microgrooves. That actually reduces the surface contact to the valve casing - speeding them up.
As far as speed is concerned, it is insignificant if the spring is on the bottom or top. I suspect that the top sprung valves are more reliable because the rest of the parts are of better manufacture - not the location of the spring.
The biggest factor in speed is the position of the hand/fingers in relation to the valves. If we play with the tips of our fingers and have them curled like holding a baseball, we have the most efficient motion and we are working WITH the geometry. If we play on the finger joints, we are pushing the valves down at a sharp angle AGAINST the designed geometry. That raises the resistance to motion and increases wear.
The amount of wear also plays a big factor in the speed of the valve. After the valve is "broken in" the wear pattern is complementary to your hand position and it should be fastest. Over time, more wear will change tolerances making oiling more critical for optimal performance.
The last point could be made for the Selmer Radial 2 Degrees valve cluster. The geometry more closely matches our hands but the manufacturing costs are MUCH higher.
Alan's Trumpet Page
or
Histoire - De 1940 Ã* 1979 - 1968, les trompettes modèle "Radial 2°"
I have a D/Eb Radial 2 that I bought new in the 70s. It is still AWESOME. I wish this was still available!