In comments on the new game, a lot of people have been complaining that the accuracy calculations feel wrong or unfair. I wish this was new to me, but exactly the same thing happened on the release of both of the previous games.

In the interests of science, I added a new function to my editor. It simulates 10,000 attacks using the actual functions from the game (the editor is the game, compiled with different flags). Here are the results:

 

So. That’s that, then.