So after compiling EU leaderboard stats a couple of weeks ago I decided to take a look at paragon distribution of banned vs not banned players. https://drive.google.com/file/d/0B1neYXz0zXCNTkM1eE5iT1BfSVU/view?usp=sharing
Here's the distribution of plvl (at the moment of rift completion) among top500 spots in each solo leaderboard
Here's one for top100 spots: https://drive.google.com/file/d/0B1neYXz0zXCNOHJqczBnLVBKZ1U/view?usp=sharing
Ofc there's an indication bias because players that grind paragon like crazy are also more likely to bot than those that don't. And there's sampling bias because those players are more likely to attract Blizzard's attention.
For many classes, there's a clear divide between 75% percentile of non-banned players and 25% percentile of banned ones. This makes sense for Wizard/Barb as they also seem most paragon-dependent for GR pushing. DH has a really big variance among plevels of banned players. WD in top100 has surprisingly small difference in plvl between banned/non-banned.
Here's the distribution of ranks among top500 spots:https://drive.google.com/file/d/0B1neYXz0zXCNMVBKSTZKcnZfalE/view?usp=sharing
And here's one for top100: https://drive.google.com/file/d/0B1neYXz0zXCNeTREQjdUVVdpQUk/view?usp=sharing
Monks and Wizards seem to experience the most top-centric bans (also Monks have 3 outlier bans very close to 500s). WD in top100 are the only class for which median rank among banned players is lower than the median rank of non-banned players.
Here's some numeric stats over banned players:
The most banned class are Wizards (surprise, surprise!), the least ones are Monks and Barbs. Overall, 155 spots out of 3000 (6 boards, 500 spots) are banned, which is 5%. Of the top100, that's 71 spot out of 600, which is 11.8%.
Any feedback/suggestions for further analysis are welcome :)
EDIT: added rank graphs as well.