This is why I'm concerned for the next phase of the MMR.

The new MMR isn't based on rank at all, you can't just "replace rank 14 with MMR 1000" because they're not measuring things the same way. Also the new MMR isn't reset every month like the current system so that part of the complaint is moot under it as well.

Also keep in mind that even with a solid matchmaking evaluation the game still has a time limit it imposes on itself to match people up. If you're a medium skill killer but the only survivors available are higher rank than you then you're going to be matched up with them rather than keep waiting for a perfect match. And because half of the games involve 2-, 3- or 4-person swfs you'll always have high skill players in swfs with lower skill friends in the same game which also will cause odd looking matchups.

One thing I'm hoping for out of the new MMR is hopefully it's appropriately increasing the rating of the survivor group based on its swf composition. The more survivors in the group that are in a swf the higher the rating of the survivor group should be as a whole. That way swfs will tend to face tougher killer players than if they are all solo which will help balance out game results between those two types of groups.

And of course hopefully they've worked out what cause the new MMR to be putting streamers with thousands of hours up against literally brand new players. (My guess is it was some glitch in how they calculated the initial seedings but who knows?)

A lot of confirmation bias in this thread, as well as lack of understanding on how probability works.

People look at the "4%" and decide that is basically an impossible chance to escape. But its not. If a survivor tries to escape off the hook they get 3 chances, so the actual odds are calculated as:

1 - (0.96^3) = 11.5% rounded up, where 96% is the chance that a survivor does NOT unhook themselves. 11.5% still sounds really low to our human brains, but its actually around the same as 1 out of 9 chances, which is low but not unreasonably so. If I play around 20 games as killer and the last survivor tries to escape the hook, I could expect around 2-3 times they will actually succeed.

There is also the fact that this chance is only for ONE survivor escaping. But as a killer you face 4 survivors who can all attempt to kobe in a match, Now its unrealistic to assume that all survivors will try to pull themselves off the hook every game, so lets examine one of the more likely scenarios that at least one survivor would attempt this:

2 survivors are dead, 1 is on their first hook and the final one is getting chased. the hooked survivor attempts to escape and the last survivor might do the same if they get caught:

1 - (0.885^2) = 21.7%. That's the chance at least one of those survivors will escape the hook, around the same as 1 out 5, a little bit higher. So if I play say 40 games as a killer and end up in this situation where I'm about to win and get a 4K, I could expect at least around 8 - 9 times one survivor unhooks themselves and might get escape alive.

I think I've made my point pretty clear. Yes its annoying to lose a survivor due to what seems like pure luck, but keep in mind that you are more likely to encounter survivors attempting a kobe as a Killer in general, simply because there are more of them that can attempt it. The game isn't rigged, that's just how the laws of math work in our world.

One small flaw in your analysis is that every failure immediately reduces the hook timer and many players intentionally don’t attempt an escape to maximize their time in phase 1 to allow allies more opportunity for a rescue. You can’t attempt to escape three times in a row unless you do it relatively quickly after being hooked, if you wait and try to escape you’ll only get one or two chances before being knocked to stage 2 if you fail.

So the only survivors that actually potentially use all three chances to escape are the ones that either know their allies aren’t coming for a rescue (because they already escaped or died for instance) or that simply don’t care if they fail or, occasionally, because they have a perk or something that boosts their escape chances. Otherwise the most common scenario is the survivor doesn’t try to escape until possibly the end of stage 1 and only does it once.

My biggest issue with MMR is that survivors and killers often lose to map RNG. It's so wild to me that we're leaving an RNG system in place that often leaves half of some maps completely barren or completely loads them with jungle gyms while also implementing MMR. Really putting the cart before the horse. What are the survivors supposed to do when the killer gets a 4 gen on one side of Azarov's with one pallet on half of that end of the dogbone? Let's make sure we tabulate MMR on that one.

I understand perfectly well that most survivors don't try to immediately escape. I was simply highlighting the fact that the odds of a survivor escaping and affecting your match as a killer is significantly higher then people realise. I already stated that the likely scenario for people to attempt this is when a game is pretty much unwinnable due to a killer managing to kill at least one or two survivors early, increasing the likelihood that at least one of the two remaining survivors will attempt to kobe.

The point of my post was to try to show that the odds of any survivor escaping is significantly higher than it appears to our human brains, simply due to how math works.