I'm still a little weirded out by pounded banks. The shortening of the angle seems stronger than can be explained by just a lack of natural roll (as compared to a 'regular' bank). I'd expect a sliding pounded ball to come off the rail close to the mirror angle, but it seems like it's coming off much tighter than the angle in. If that's not due to rail friction, what is it?
I've been playing under the same assumption as you, Creedo. I've always felt like it was possible to shorten up banks much more than that mirror angle using only speed. By how far seems to vary with the playing conditions.
Another reason I say this is, the forward roll should have less - maybe even negligible - effect for banks that are closer to perpendicular to the rail. It still seems like you can shorten those up a considerable amount using only speed.
It should be fairly easy to test whether lack of forward roll is the only factor shortening up those banks. With the object ball close to the rail (but not touching, so as to eliminate as many variables as possible), there should be a range of shot speeds that will result in the object ball sliding into the rail. If the rebound angle changes with shot speed, then there's got to be another factor involved.
Maybe someone with the time to play around with it a little can help us out with the testing. I won't have a chance to look at it until this weekend.