modu,
I think I understand what you are getting at. See if this helps.
===
You understand that each atom is very transparent to gravitons. Most of them pass through a given atom as if it did not exist, but every once in a while a graviton will hit something inside the atom. This gives the atom a push.
You also understand that, despite the low probability of such a hit, there are so many gravitons passing through a given atom that each atom gets hit a lot (trillions? of times per second, billions-of-trillions? – we do not know for sure, but it is a lot.)
You further understand that to a first approximation (and to a second or third approximation), each atom in a larger group of atoms (such as a cylinder) receivers an equal number of graviton hits from all directions.
But –
You also see that, for gravitons moving in the vertical direction, the probability of a graviton hit for two atoms aligned vertically is greater than the probability of a graviton hit for the same two atoms aligned horizontally.
===
I believe you are correct. However, this effect it is too small (in theory and in reality) to be detected with state of the art instruents. Atoms and the things they are made of are very very very (did I mention very?) transpartent to gravitons. Stacking them has no effect that is detectable (at this time).
If you and I are correct, some day this effect will be detectable. At that time the effect will probably cause many scientists to slap their forhead and say “doh”.
Sigh. Too bad we are not there yet.
Regards,
LB
