I believe that usually truth lies in the middle, not on the extreme ends of opposite sides.
We cannot disregard easily neither bet selections, not bet progressions.
The thing is that a large bankroll of say 100,000 $ or even more is harder to broke but the persons who are gambling with such bankrolls are not betting the minimum, their initial bet is something like 25,50,100 or even more per unit, therefore whether you have 1000 $ bankroll and bet 1 $ units or 100,000 $ and bet 100 $ units it's the same in terms of risk of ruin.
It's highly unattractive for 100,000 bankroll to bet with 1, 5 or even 10 $ units, additionally is needed sufficient spread between minimum and maximum bet limits.
A high maximum with high minimum is no good, neither is a low minimum with low maximum, the greater the difference between minimum and max bet the better.
Another fact is that gamblers have lost millions at roulette tables, if the brute force of vast capital was sufficient to guarantee profit, that would never happen.
Imagine bankroll as force, imagine prediction accuracy as technique, the best warrior must have both, force and technique.
There are several examples in history where the larger army has been defeated by much fewer men, also in one on one combats smaller individuals beat larger because they used their technique, their intelligence to beat brute force.
A few examples; David and Goliath, Leonidas and the 300, Trojan horse, Mongols were many but this was not the key to their largest unified empire ever existed, they were very disciplined, with well organized and extensive scouting teams and formidable archers.
In other a combination of things makes the best possible results, we cannot rely only in one asset.
It's wrong to consider that since casino has much larger bank than ours it's going to win at the end because the aim of every gambler is not to empty the entire bank of the casino, but to win some money.
If someone was trying to broke the bank, then yes, he/she would be in great disadvantage if he/she wouldn't possess at least equal amount of bankroll.
Another misconception is that time is against the players, what matters is not the duration but what happens on that duration.
I'm sure you have experienced shorter and longer sessions with positive and negative results, time doesn't control the outcomes but gives the opportunity to every possibility to happen.
So if you start winning and decide to go, the next time you'll go will be the continuation of your previous time and every time thereafter will be a continuation of the first time.
If you think that every time is not continuation, then think the following, if you could win easily and stop while ahead equally easy will be to begin negative and the what would you do?
Take the loss and go? Or risk more money to win less?
So as you see, neither hit and run could help anyone.
My philosophy is the following, every bet is a part of a session and every session is part of the long term, thus if we could win most of our bets while maintaining the balance between risk and reward, we would become long term winners.
Win ratio could be deceptive because it risks more money to win more times, this is not the solution.
Positive progressions increase after wins, negative progressions increase after losses but do wins and losses follow a specific distribution pattern??
Instead of focusing on the distribution, it would be more accurate to focus on totals and establish a sophisticated money management on that basis.
Every bet selection has each unique characteristics, therefore a progression should accommodate the specific characteristics of that bet selection.
First of all is important to understand a few things like, why my selection wins/loses, why my method has the potential to become long term winner, why I would bet this instead of the other.
All these questions will help to determine which bet selection has the best possibilities to be long term winner.
House edge cannot prevent long term winners, this doesn't mean that house edge doesn't exist nor that the casinos will still have profit.
Let me give you a simple example, I'm betting always on Red and after 100,000 bets Red has appeared 50,900 times while Black and 0s were 49,100 times, this means that I've won 1,800 times more than I lost.
These figures are really possible and nobody can say with certainty that after 100,000 results would be 48,650 for one side, 48,650 for the other and 2,700 for 0.
Actually what happens in reality is not this balance but smaller or larger deviations, the larger the total amount of results the larger could be this deviation.
Within a small sample like 10 spins is much more likely to encounter a balance, but think about 100 spins sample, 1000 spins sample, apparently becomes more and more possible to deviate from the theoretical mean.
Expecting results to balance is VERY big assumption, actually reminds me of the gamblers' fallacy, what could happen within 100 bets is just a microcosmos of what could happen in 100,000 bets, the same in larger scale.
You could find a streak of 20 EC within 100 and 100,000 but in the larger sample chances to find more than 1 such streak are tenfold.