Woof... Man... Seriously... Not trying to rain on your parade...
But in a nutshell dude, you scare the crap out of me.
There's so many things to address, but I'll keep it simple and to the point for now, and go after the worst.
1. I hope it's just your use of the terminology but the P and the S are *not* interchangeable in their order. 5S9P is *not* the same as 9P5S, they represent two radically different animals, despite having identical characteristics at the main terminals. I hope for your sake you're actually employing the latter.
2. 70mV is a LOT of imbalance cell to cell... Toward the middle of the SoC curve this represents not only non-trivial but *significant* differences.
3. #2 is not surprising, as you would appear to have randomly mixed cells from countless sources, with different ages, cycle histories, and other critical considerations such as abuse (at either end) and so on. This is an annoyance with most battery chemistries. It's downright dangerous with Lithium NMC.
4. YouTube doesn't capture the potential when a cell gets a hair and decides it doesn't want to play nice anymore. Ask a buddy of mine that once viewed them in the more lax sense you appear to. Somehow he managed to get an "annoyed" RC pack outside before it took anything with it, but it certainly wised him up on handling and other important conditions like maximum and minimum voltages, what balancing (or lack of) can really imply, and what an irritated lithium cell is capable of. He didn't "hear" what YouTube had to offer on it prior to his experience. It took an up close personal event to teach him the lessons that a screen simply can't capture. Ask him if you don't believe me, I'm sure he'll be glad to tell you the story. That was 3 series cells. Imagine what you've got going getting irked.
I *strongly* encourage you to take a single cell, place it in a completely safe environment, and push it to the limits as a test to demonstrate in person what actually happens when these cells get mad. Make sure you catch a whiff of the fumes as well... That's a fair part of the experience. Then remember that neighboring cells within a certain distance will be affected by intense heat, causing a chain reaction. Trust me, it's not something you want to deal with on a large scale level unprepared.
I don't want you to take me wrong. I personally *love* lithium... But the NMC variety (most common 18650 cells) are nothing like any other chemistry in terms of their demand for respect. I can't emphasize this enough. They're a beautiful thing, with a lot of potential. Unfortunately, for as much bad as good, when mistreated.
Voltage limits (particularly the lower end) tend to be taken "grain of salt"... With overcharge getting the majority of the attention for someone just going in. When in fact while over voltage can and will indeed set one off, overdischarge is the monster lurking in the shadows... The cells change (dendrites namely) in overdischarge conditions, leading to a short at some point in the future... This unknown is primarily why BMS systems *permanently* disconnect a pack from the terminals in commercial products... It's simply *not safe* to EVER charge an overdischarged cell, ever again.
Just looking out for you, as well as your readers. I'm all about the experimentation... But please... Be safe about it. This protection and balancing stuff should *never* be any kind of "afterthought"... Not with these...
Carry on man... Be safe.