If you are referring about roulette my best advice is to test your strategy separately onto different specific single generators.
No one general strategy can work on every wheel of the universe, otherwise roulettes would have taken off from the gambling world.
The more the factors involved are uniform, at least taken by a theorical point of view, the best will be your estimate on how good your system will fare on long terms.
Imo, random world works on multiple levels:
level 1 is whenever a "more likely event" (having a p>50%) will occur at the same or almost the same occurence dictated by general probability laws with a low impact of variance.
level 2 comes out whenever a more likely event presents a moderate impact of variance.
level 3 is when the variance is so high and frequent that it will easily counter act any strategy.
A perfect scenario will be to get a lot of #1 level results and some level #2 outcomes with no or few #3 levels.
Since we cannot rid of those unfavourable #3 high varaince situations, we must find ways to restrict their impact over our strategy.
Say we have a plan to win whenever every level #1 and #2 will come out along the way with no #3 level situations.
We'll be millionaires.
Unfortunately certain unlikely level #3 situations will come out along the way.
Chasing them is a big mistake.
Level #3 events will erase every profit coming from #1 and #2 situations.
So we are forced to stop our betting anytime a situation #1 and #2 is surpassed.
Therefore we must classify many results taken from the same wheel in order to get a picture of #1, #2 and #3 events impact.
The more #1 and #2 events any wheel will present over the long run, the best will be our results.
As long as results are taken from the same parameters involved in the same wheel, things cannot change.
Of course humans hugely interfere with such process so let's guess which wheels I'm referring to.
A profitable wheel is any wheel capable to produce an unproportional higher amount of #1 and #2 situations as opposed to #3 events.
Such #3 situations will happen anyway but at a lower degree.
So getting a list of which wheels are producing the best #1-#2 vs #3 ratio is the best tool to get an edge.
Just to make an example, the Caesar's Palace IB automated wheel (placed in proximity to the Omnia nightclub) perfectly fits such requisites.