The problem with statistics relating to DBD and BHVRs intended balance
I'm sure no-one will say that statistics are a completely useless source of imformation, however; the devs seem to rely on them too much.
An example of this is the ruin change. The devs said they nerfed it because they saw it was being used in 80% of red rank games and they then saw this as the perk being too strong. They never actually looked at why is was being used so frequently. The answer was that it was almost necessary on low-tier killers to be able to use AND play nice (no tunneling, slugging etc). Now these killers have to basically use PGTW/Corrupt and play 'dirty' to stand a chance against good survivors.
I'm not 100% on this, but I believe BHVR also said they were happy with legion because he wasn't underperforming according to the kill rate. If this is true, why don't they buff nurse as according to the statistics she is the weakest killer in the game by far. The answer to that is that BHVR obviously know that the kill rate doesn't reflect the actual rankings and strength of killers and they saw this here.
Part 2: The Devs Intended Balance
It has been stated by BHVR that they plan to balance the game around 2 kills and 2 escapes. However, most of the times you get 2 kills and 2 escapes you will rank down as a killer. Also, playing killer at R1 (ignoring MMR for now) I get far more 4Ks or 4-man escapes than balanced 2-man escapes 2 kills. This says alot about the balance of the game; Either the killer destroys the survivor or the killer gets destroyed.