My playing experience and testing has taught me that things don't often go 3 times over their expected appearance. A dozen/column should average a hit every 3 spins. So I think the best way personally would be to wait for 9 misses and then hit it for a few spins. If it still doesn't appear, then there is a chance that it's going to be one of those long sleepers that you don't really want to be chasing.
This is true for every group. It is the solid foundation needed, to build systems that exploit what is missing.
But a lot of homework is required. You need to have an idea as to what the optimum length of a streak is
before it's interrupted, (most of the time), and the anticipated result shows up. It can't be too long to save time. And it can't be too short, to overpower the bets that will follow.
Just optimum. A little homework can determine that with excellent statistical accuracy.
Then you need to research for the ideal number of spins that guarantee positive results the majority of the time. You don't need a long progression with its perils. And it can't be too short a progression, that it is unrealistic to succeed. It has to be just right. We call that the "the most frequent winning range".
It is far more important to know the most suitable length of a streak, before it discontinues, than knowing the extreme length a streak can reach.
Then the size of the Bank Roll has to be taken into account too. Waiting for 9 missing spins for a dozen and betting with $1 after that, won't be very productive. Too long wait for the benefit you will get. Maybe it has to be scaled down to 4 or 5 misses.
But for the player who bets with $50 and $100 chips, the wait for a long streak is well worth it. Big benefits, with higher certainty.
And last but not least is the use of virtual losses. That is an option that you can use , if you are the patient type.