phisical forces... (emulating the reality) VR ENGINE FOR MSX!!!!!!

come on.... nobody is a phisician?

Well, I am a physicist. I've done quite some modelling as well (ocean currents).
However, I'm afraid you're utterly over-ambitious. Unless you make some heavy simplifications even simple classical mechanics will need loads of processing power which MSX does not have. Don't even think about friction. That's nonlinear stuff.

Still, you might use simple things like conservation of momentum.

The exact "impact energy" spended on the crash of two objects taking in account the forces kinetics , forces centrifugal, wheight of the object...
This confuses me a lot. What do you mean by "forces kinetics"?

If one denotes the total force by F, F is just the sum of all forces on the object (like centrifugal, coriolis forces, friction and gravity.)

but also that depends of the weight of each object ....
Not only the weight. You also need to know what each object is made of (matter) and how that type of matter responds to an impact. Intricate stuff.

Let's forget about impact and focus on dynamics only for now.

Conservation of momentum

A ------> <---------- B

Suppose we have two objects moving in 1 dimension (along an x-axis). Object A has a weight m1 and velocity v1. B has a weight m2 and velocity v2.

The momentum of A is now given by p1 = m1 * v1.
The momentum of B is equal to p2 = m2 * v2.

The law of conservation of momentum states that p1 + p2 before impact is equal to p1 + p2 after impact.

p1 + p2 = p3 + p4 (p3 = momentum of A after impact, p4 = momentum of B after impact).

This way you're able to calculate the velocities after a collision.
However, when damage occurs momentum will get lost and hence this law isn't valid (it's only valid when all momentum is exchanged after the collision)

It's no problem at all to expand the above to 2 or 3 dimensions. The only difference is that p1 and p2 are now vectors (well, the velocities are and thus the momentum as momentum = weigth times velocity.)

For instance, p1 = m1 * (u1, v1)
where m1 is the mass, u1 the velocity parallel to the x-axis and v1 the velocity parallel to the y-axis (the speed is then given by SQRT(u1*u1 + v1*v2).)

p1 + p2 = p3 + p4 then leads to an extra equation:

m1 * u1 + m2 * u2 = m1 * u3 + m2 * u4
m1 * v1 + m2 * v2 = m1 * v3 + m2 * v4
(note that m1 and m2 do not change after a collision while playing pool, those values are fixed. In particle physics it would be possible though.)

And as you usually only knows the speed of a ball you need to calculate it's 2 components u and v relative to a base (normally your base is just a simple x-y plane).

Example.
A ball travels with 1 m/s and it's direction is 30 degrees anti-clockwise.

y    velocity vector
|   /
|  /
| /
|/  \ 30 degrees
---------- x

Hence its x component (u) equals speed times cosine (30 degrees) and its y component (v) equals speed times sine (30 degrees).

from this example i not figure out how to calculate the corrects directions taken according with

if the balls craches fully ... by its centers.....

or if the balls crashes by a side or just touches a bit.....