This roulette strategy lives up to its name "Zufall" (German for "coincidence").
The concept follows the idea of overcoming coincidence with coincidence.

Two simple chances are chosen at random
and the stakes get optimized to their difference rate if there are opposing opportunities.
(It would make no sense to put 16 pieces on black and 32 pieces on red at the same time.)
It always starts with 16 pieces per chance.
Then the bet is doubled on one chance and halved on the other.
This goes on as long as the opening balance ( on balance) is not exceeded,
in the event of a balance win, the whole procedure starts again from the beginning.

trizero:

The concept follows the idea of overcoming coincidence with coincidence.

Clever idea! That zufall ;-)

The math here speaks for itself.

With 16 pieces on red and 16 on black, the expected value per spin is:
2 × 16 × (18/38) - 2 × 16 = -1.684 pieces

No matter how the bets are redistributed between the two chances, the house edge remains constant at 2/38 = 5.26 %.
Doubling one side and halving the other does not change the total amount wagered per spin -
it only shifts the distribution. The net loss rate stays the same.

The "optimization" via difference betting reduces variance but cannot affect the expected value.
Over N spins the cumulative loss converges to N × 1.684 pieces regardless of how the progression resets.