ratings stats distribution graph glitch?

And you should read the paper describing how Glicko-2 works (https://www.glicko.net/glicko/glicko2.pdf). There are no "certain parameters in rating calculation" that would "change at every multiple of 100" - and no "k factor" that would change at certain steps in ratings. The rating updates are determined only by

(1) game result
(2) the difference between opponents' ratings
(3) their rating deviations

And rating deviations are determined by previous results so no computational artefacts or steps there either.

Read the discussion on statistics stackexchange if you are capable of reading and understanding.
Are these attacks really necessary? Are you unable of decent communication just because people dare to disagree with you?

Well, you have been given en easily understandable explanation.

Also, the source code is open source. You can easily check yourself on github. You will see that there is nothing special going on. No special pairing. No pairing due to winning streaks. No pairing due to imaginary milestones. No nothing.

Please stop spreading those idiotic conspiracy theories, and accept it for what it is: a relic of the human tendency to take a break at those points. And when people stay longer at those marks, they cannot be on the neighbouring levels at the same time. So from all surrounding rating-buckets a small fraction is now shaved off and is on those round number spots instead. This results in the spike-building around those x00 marks.