And then cheating rises again.

So whatever caused cheating to decrease relative to the population (assuming that's what's being measured) is probably to blame.

Here's a different explanation:

1. EA announces new update, so users flock back before it goes live (first major dip)
2. Update goes live (steeper dip) and more users return to try it out; some cheat engines don't work on the update, which adds to the dip
3. Cheat engines start getting patched, hence the rise after bottoming out; number is still lower due to larger population (same number of cheaters)

A better test is to keep support for Linux for the launch, then drop it in a minor patch update a few days later.