Unusual distribution of steps needed to reach Kaprekar's constant (6174)

In the distribution of the number of steps needed to reach Kaprekar's constant (6174), I observe an unexpected distribution pattern, with three steps being the most common number of steps required. I cannot think of why this is the case. Has anyone done some though about this phenomenon? See a plot here: https://earth.hoyd.net/posts/distribusjonen-av-antall-steg-til-kaprekars-konstant/