If they simplify the assumptions to the point they have left reality I think any conclusions are worthless. Even a super-ball is under 95% COR and the rubber in the rail is nothing like that.
Your interpretation seems to be incorrect and 0.55 that you reported is in error.
Please see the following article which
emperically uses high speed cameras to measure ball speed, and the excerpt from page 791:
From Fig. 9 we see that the relation between the rebound
and incident speeds is almost linear for the incident velocity
in the range of 0.28–3.5 m/s the typical range of ball velocities
in the game. A best fit straight line for the reboundincident
speed data gives a coefficient of restitution of
0.818
for this velocity range. The results are more closely fit by the
second-order polynomial y=−0.0877x2+1.131x−0.0953,
where x is the incident velocity and y is the rebound velocity.
These results are not valid for a general ball-cushion impulse
but are applicable only under the conditions of no-sidespin
and pure rolling motion prior to the impulse. We believe that
the ideal variation between the rebound and incident speeds
should be linear and the reduction in the coefficient of restitution
at higher incident speeds is due to cushion deformation.
The gradient of the plot at
lower incident speeds is
around
0.910, and this value shall be valid under the assumption
of a rigid cushion.
Marlow6 reported that the coefficient of restitution for rails
in a billiard table is 0.55 but did not give much detail about
the experimental procedure. He compared his results with the
values suggested by Coriolis1 and concluded that they agree
closely.6 The cushion height for snooker is 36 mm, with the
ball radius equal to 26 mm, which is close to the height of
1.4 times ball radius found in pool. Thus the cushion and
ball geometry is almost identical in pool and snooker. It is
possible that Marlow considered the rebound ball velocity
at
the end of the sliding phase rather than
the correct one immediately
after the impulse. Then the coefficient of restitution
for the shot could be 0.63,
but this result has no physical
meaning.
