Here's what I remember thinking about when it was being discussed. We
all know if we skid into a turn using only rudder the ball swings to
the outside, and if we roll into it with ailerons the ball rolls
inside, so with the correct combination (since we are all expert
pilots, it's always like this for us) the ball is nailed to the center.
However, we will, if in level flight, feel increased g's -1.41 times
body weight at 45 degrees in a coordinated turn, you guys all know
that. However, if you while coordinating rudder and aileron you put
forward pressure on the yoke, you can take off those G's -- push hard
enough and you can probably go to negative g's and only have the seat
belt holding you down. So, there is some forward pressure that will
just compensate for the additional g's the coordinated level turn will
cause.

So, I can talk myself into the start of a roll (more barrel than
anything else) thinking I can keep the ball centered and the scale I'm
sitting on registering my actual weight. The engineering analysis can
continue from that point, saying at any incremental distance or roll
angle there's some combination of aileron, rudder, and yoke that will
result in an additional roll with one g postive force into the seat.

I remember someone running the numbers and concluding the diameter of
this descending barrel roll thing would be about 80 feet (not far from
a snap roll, huh?). That's nothing the little rudder and ailerons on
airplanes I know can do. The question is, is there a simulator jock out
there who can call up the parameters of a real life airplane that can
fly the thing? Do simulators allow the inclusion of a three axis G
force meter on the panel?