More importantly though, I don't even know what telemetry is telling PGI... whether it be the massive reduction in use of most upscaled Lights or the semi-random nerfing of quirks, particularly on the IS side in compensation of the upcoming skill tree, but it's certainly seemingly having no rhyme or reason.
To break it down, this will be discussed in 3 parts:
1) Engine Desync - Did you really do that?
2) Telemetry - Is there a reason that the IS needs to be screwed even more?
3) Balance - Do I need a 20-sided dice for this?
1) Engine Desync - So a KDK-3, Atlas, and King Crab
I like being a numbers guy... at least comparing X stat on Y mech to another... because you know... curiosity and all that fun stuff. While I'm sure the title is a dead giveaway to an obvious blunder or intentional design... I felt like the fundamental idea of what engine desync was not understood.
The point of it is to make sure that mechs that needed more agility, but wouldn't normally gain it due to having too small an engine cap to get it. In some cases, mechs that were too fast yet were not Lights would have a slight nerf. Of course all of this would be still be a case by case basis.
One of the most notable changes is actually something we commonly see in many matches... the Hunchback-IIC.
It's agility is actually equivalent to a Highlander-IIC (also equivalent to an Archer, Cataphract, Warhammer, and Battlemaster).
Mind you, this is a serious nerf and while I feel it may still get used, it's not going to be an attractive option.
If PGI feels that is the way to nerf things... well, good luck with that. It's going to be forgotten like many other IS mechs, such as the Firestarter, Jenner, Shadowhawk, Stalker, and many others that people forget over time... while many of them used to be really good at one time or another (with or w/o quirks).
There are plenty of other examples as well, but those are the most egregious ones that I can see (outside of the obvious Spreadsheet errors - it's not Excel's fault the Viper-C got "screwed").
2) Telemetry - 5% use in Tier 1 is not the same as 5% use in Tier 5.
I would call what I've just mentioned "unquirk to forget". Basically, any mech that was once a meta any point was routinely ignored over time, assumed to still "be good" (from unreliable sources) when it actually it had slowly fallen off the face of regular play (let alone comp play). Often this is mentioned in mock tones, because most often random rants about "it was good once upon a time" was dependent on two specific related responses. Normally you'd want "it was nerfed and now it's not stupid crazy OP"... instead of "if was nerfed, and people stopped using it altogether". We get a lot more of the latter than the former.
It's harder to notice when you're not playing against actual competition to test your understanding of the changes. While I'm not saying comp is always the definitive answer to everything, but patterns start from the top. There will always be exceptions (like the "Mist Lynx Whisperer"), but thing about the top is that if something... whether it is weapon or mech that becomes "unusable", it stops getting used as quickly as possible. The trickle down effect is exhibited months later (for whatever strange reason) where it becomes readily apparent.
It's a really bad habit when we continue to hold a lot of the same cycle of same bad mechs that never get attention... and ironically most of it is IS. New tech will not fix everything... it's just a bandaid over a bigger problem until it is fundamentally and properly addressed. Stuff like LFE is only going to truly replace STD engines. XL is still where the action is truly at, despite the obvious cost to compete. That's before we come to the real issue when the Skill Tree drops by...
3) Balance - 91 Clicks Per Mech/Ain't Got Time For Dat
This is something people have to remember... the skill tree like the old one applies to every mech. This means that every Clan mech that had zero quirks is getting a buff. So many IS mechs are actually getting nerfed in anticipation for this, but it could actually easily be argued that the preemptive strike isn't as warranted.
The most notable example is the Grasshopper-5P. The nerfs feel a lot similar to the rescale/Black Knight nerfs of yesteryear. Black Knights are not really a thing these days (primarily due to said nerfs, but the upscaling didn't help). Grasshopper-5P runs a niche that most heavier (let alone IS) mechs do not run... that is actual long range laser vomit. Clan ERLL is very onerous to most players (the Supernova that has the ERLL quirk is actually very devastating, but is also getting nerfed in the upcoming patch based on the Clan quirk notes). While yes you can get duration quirks on the tree (for which I don't know what the reasonable # of nodes is required for this), the Grasshopper-5P is going to be a lot less practical for all intents and purposes (I mean, ISERLL requires quirks to be semi-useful vs PPC+Gauss). High mounts are still a thing and are valuable... but any long term staring at a more useful/prominent mech like a PPC+Gauss poptart is very hazardous and makes this mech a lot less desirable (it's already niche as it is).
To understand the game, you must understand the context of why a mech functions the way it does at the highest level. Versatility of a mech is all about where its mounts are and its quirks. Failure to have that will produce "functional but not optimal" mechs. Most of the time though, having a niche is better than not being relevant at all. If only people truly understood what the rescaling did for Lights and what the actual problems were (not perceived, because a lot of bad players are often caught in bad positions and would've been punished regardless of mech), we wouldn't continue to see most upscaled Lights not even be considered in most instances (Jenners, Firestarters, Panthers - some of which are getting additional nerfs in the next patch).
TL;DR
When you don't leave your data to peer review (as in, feedback from players that have a clue), flaws in your "experiment" continue to be revealed. It doesn't take much effort to ask for help... but it's a lot of effort in learning, understanding, and articulating issues that plague the system. Balance is not always something you can chuck numbers at, but also you have to actively playtest in a live environment and get continuous feedback for. When that doesn't happen... well, we have plenty of time to gloss over and mock it.
When you're not paying attention, eventually errors in your design will catch up to you and affect everyone.