I am no fan of a method that brute forces a net win in one good hit. I consider progressions like Martingale a failure too, considering the probabilities that could show in successive bets in roulette or any other game frustrating such ideas as fools' gold.
  However, such ideas could yield us with 99.999% accuracy(I leave 0.001% for exceptional cases that might creep in without asking for our permission even in our first attempt), if and only if we attack the virtual limits of a bet. What is the virtual limit of a non hitting dozen is debatable too but I consider 45 spins as the virtual limit.
                                                                      Now, waiting for a dozen numbers to not hit about 37 spins or more in a stretch will be insane but still doable, in a particular way, that I will explain later. How about brute forcing those 12 numbers for a single hit in first to 8th bet maximum?

those 12 numbers about a third will come out in the next cycle.....the better option would maybe a better progression that would encapture that 3rd of them ...so a profit on first hit may not be possible but a progression over 4 hits in next cycle spreading the cost may be better ..but is that possible?...

Quotethose 12 numbers about a third will come out in the next cycle.

You mean to say that in next 37 spins only 4 of them are most likely to hit while 8 remain unhit?

more may hit but 3 will be guaranteed..4 is the average

In my observations, about 6 new unhit numbers show easily from the remaining in the next cycle of 37 spins. Exceptions do not become the rule but as a cautious approach, even if we take 3 nbrs to hit, there need not be only single hits of them. If we bet all of these 12 nbrs as our fixed bet, we have to get minimum 6 to maximum 15 hits in the coming cycle. We have lots of database of real casinos plus random.org to analyse the same.

                    If we take minimum 6(as I assert) to be true, is it possible to brute force a confirm win?

i can only say try it out...the 6 hit is not guaranteed ..doesn,t work with the mechanics of roulette...that's 50percent of what's left..but you do get repeats on unhit numbers too that's true ..but again..that's included in the laws...12 unhit and above are best to play..but not going for half of unhits..you can a lot of the time get more hits than average i don,t disagree..but anything needs to be based on a third..this is the reason it takes so many cycles to hit all numbers..you can get the one number in the end sleeping for a long time..but by that point you would not be playing..every spin is the start of a cycle..and cycles overlap...
but as i say try it out..random .org spins are flawed..real data is the way to go

I discard this way of playing. It is as bad as playing simple martingale.