Can you combine physics and roulette systems methodology ?
That is a interesting question.

Lets assume we take advantage of the random parameters and back engineer things.
When you deal with visual ballistic you predict where the ball will land, the high probability area.
But lets say we reverse things and play the negative area and skip the high probability area (where the ball should end up)

This is what you can do with physics.
You can measuring the ball in the beginning of spin and know how many turnarounds the ball will make on the ball track (before hitting deflector).
So lets say we only play when ball will make 17 turnarounds on the ball track before hitting a deflector.
And we will use the same rotor speed each time.

So ball will travel constant between A to B with the same time-frame.
And same with rotor.
This means that the distance with both will be same and the only random factor left is the deflectors hits.

Now if one and same deflector would hit all the time, then the ball would have pretty much the same traveling distance for each spin.
And we would have our high probability area where the ball would end up.
Could be a sector around 6 to 12 numbers.

Reverse:
But now assume we pick a level wheel where all deflectors hits evenly.
Then ball has to start hitting one and same deflector (repeating several times) to have a chance hitting the sector we don't play.
And even if, so will scatter make the ball miss sometimes.

So you estimate the ball and rotor speed so they don't become random element.
You only play for example the spins that make 17 turnarounds with one specific rotor speed with random hitting deflectors.
So the only random factor you count on is the deflector hits.

After collecting data without betting you could see if you need to place 15 to 24 bets.
Measuring variance.

Hi, Wheel,
maybe this will help..
http://betselection.cc/general-discussion/kimo-li%27s-%27edge%27-quote-maybe-i-can-try-to-explain/

thanks..