Roulette Ideas

Let's suppose that House Edge is an insignificant issue, merely a taxation on our profits, thus the major obstacle we have to overcome is variance.
So how could we deal with variance?

For the example's sake I'm going to present you with a hypothesis on which we are gambling in a 50/50 game, such as even chances on Roulette, "banker" VS "player" on Baccarat, "pass" VS "don't pass"...etc.

When there's positive variance we have to do nothing because we are winners, but when we have negative variance we have to wait until variance undone its own manifestation...in other words, if we were trailing by 10 losses more than wins, we would be unable to overcome the negative deficit simply because we can never be 100% sure about when exactly a win is going to occur, with any kind of progression/MM.
Yes. we are going to win eventually, but since we don't know exactly when it's useless, or even harmful to attempt to raise the bets when negative variance still in action.

So by making an assumption that negative variance will eventually anule,offset,negate its own negative with a positive variance the only thing which we would have to be concerned with is the HE in regards with the duration since the negative variance begun till the time negative variance will end.

Let me become more specific;
I've found myself 10 losses in the hole, I kept flat betting for the whole duration till after 3 cycles of 37 spins/bets the positive variance made me win 10 times more, effectively offsetting the negative variance before, but since there were 111 bets I'm still -3 units overall thanks to the House Edge, at exactly this point in time, when there's not any losses left from negative variance, we are going to bet just enough to surpass the negative balance coming from the HE exclusively!
On the specific example 4 units bet in order to come on top with only a single win in the positive side (ours).
In case that eventful moment wouldn't come with a win, then wait again until the losses will become equal to the wins, then bet again to overcome the HE deficit plus any lost units from previous, failed attempts.

I want to be 100% clear, I'm not advocating such betting method, like I said before it's merely based on the assumption that only variance could "beat" itself.
Let's assume that this hypothesis is true, but how long before setting the record straight???
Hours? Day? Weeks?

I would really appreciate your colaboration, I want to know about what do you really think, could be such an approach viable, effective and practical?