Bug in ratings distribution graphs?

In the graphs on rating distribution there are spikes going up in ratings ending in multiples of 100. You can se it for instance in the bullet graph (http://en.lichess.org/stat/rating/distribution/bullet) in ratings 1600, 1700, 1800, 1900, 2000, 2100 and 2200, in blitz (http://en.lichess.org/stat/rating/distribution/blitz) in ratings 1500, 1600, 1700, 1800, 1900, 2000, 2100 and 2200 or in classical (http://en.lichess.org/stat/rating/distribution/classical) in ratings 1500, 1600, 1700, 1800 and 1900.

I find it very suspicious and difficult to attribute to chance and randomness.

I find it more likely that there are some players who prefer to end their chess sessions only after they pass a certain rating threshold. Multiples of 100, while meaningless to Glicko2, likely hold psychological significance to many human players. It wouldn't take that large a percentage of players holding a slight bias towards reaching a rating goal to produce a chart with spikes like that.