The first few draws decide everything
A jar with one red ball and one blue. Draw one, put it back, and add another of the same colour. Do that a thousand times. You might expect the jar to stay close to half and half. In fact it is as likely to end 95% red as 50% red, and it settles on its share very early.
The rules, all of them. Take a ball out at random, note its colour, put it back, add one more of that colour. Repeat. This is George Pólya's urn (1923), and it stands for anything in which being ahead makes you more likely to get the next one: a product people buy because others have, a street that gets busier because it is busy.
Exactly: after n draws, every possible number of red balls is equally likely, 1 in (n + 1). So the final shares are spread evenly from 0 to 100%: 10% of jars end in each tenth, and the experiments confirm it. And each jar's share stops moving early. After draw 20 the share moves by under 7 points on average in the next 980 draws; a jar that is more than half red at draw 20 is still more than half red at draw 1,000 in 91 cases out of 100. Nothing about the red balls is better. The early draws were chance, and everything afterwards built on them.