I experimented with how these small structural differences can affect the output of the code with this sketch that draws generative pipes.
Here I controlled the way the pipe turns by randomizing the corners that prompt the pipe to rotate it’s direction.
Because it’s relying on complete randomness, the corners sometimes get stuck in a random loop, and keeps turning every few frames.
In another structure, I controlled the corners with the seconds on the clock.
So I generated a random number between zero to sixty,
and when that random number matched the second on the clock, the pipes would turn.
Because the probability of those two numbers matching was less frequent,
it created a structure that generated the pipes I had in mind.