The Devs when it comes to data

Alongside with Remember Me which was also nerfed due to the Obsession getting tunneled and the EGC being introduced.

My fixes for both perks:

Dying Light

You become obsessed with one Survivor.

Your Obsession gains a 25 % Action Speed bonus to unhooking and healing other Survivors. (<- nerfed from 33%)

Each time you hook a Survivor that is not the Obsession and the Obsession is alive, gain a Token.

All Survivors get a 2/2.5/3 % stack-able penalty to Repair, Healing, and Sabotage speed for each Token. (<- included the Obsession into debuff. Stacks remain when the obsession is killed like with STBFL)

The Killer can only be obsessed with one Survivor at a time.

Remember Me

Each time you hit your Obsession, you increase the opening time of the Exit Gates after all gens are repaired for all Survivors by 5 seconds up to a maximum of 20/25/30 additional seconds. (<- Old Remember Me + the condition that all gens need to be done to make it still possible for the last Surv to leave through the exit gate + included the Obsession in the debuff, so there is no need to tunnel/kill)

I just assume they have more behind the scenes than the data they provided us. They said not to draw too many conclusions since it was from matches where rift challenges were being tracked and you can't delineate red rank games to see further breakdowns or filter out "hook suicides". It would be dumb to assume the data they showed us with kill rates and pick rates is all they base their decisions off of.

Those stats surely are meaningful, especially when we consider that every 2nd match includes a DC, AFK, suicider and/or game-throwers.

They used to split the results up in ranks, are those informations also available?

It would be really nice if we actually had full transparency over such statistics - with more relevant background information - maybe then these graphes would appear a bit less.... manipulative.

Because sometimes they read data wrong.

Years ago there was a stream, where they mentioned that only 23% of all survivors use SC.

Later it tourned out, that the data was refering to perk slots.

You're right they don't... which is why kill rates are inflated

In my opinion

and how many of those kills were because of stupid team mates? There's so many factors to account for here that just looking at killer performance isn't enough.

Did the individual survivor kill themselves on hook?

Did another survivor attempt a save in front of the killer's face?

Did the other survivor unhook then hide from the killer, forcing a tunnel?

Was the survivor any good at the game? How about their team mates?

Considering how easy it is to be carried to red ranks by other survivors and how easy it is to rank up, my guess is that at least half the kills in red ranks done by killers can be attributed to survivors that don't deserve to be in red ranks to begin with. Anecdotally, I can also see this in my games, some games I'll face a team of good survivors, some it only takes 1 or 2 bad players to snowball in my favor.

They have a very flawed system in place to determine skill, and because of this fundamental failure, any information extrapolated from such a system will be flawed and will not tell them anything meaningful, most especially considering player skill and what the top players are. It's the same reason why they said OoO was fine, because it doesn't necessarily give the individual survivor more survival rate, but they failed to mention the impact on the rest of the team. OoO is certainly a skill that will favor the team over the individual as it will get the killer's attention, and also encourages tunneling the OoO survivor.

Can you elaborate on that? Why would they be inflated by excluding games with DCs?

The games with DC's are INCLUDED as kills thus inflating the kill rates