It is common for game designers to leave unbalanced aspects (cheese) in their games.
The discussion is about whether game developers nerf things in single player games
at all, and why they would do that. Leaving some instances of cheese un-nerfed intentionally in a balanced fashion, despite potentially nerfing other things, is a
separate question which is only tangentially related and not really useful in this discussion.
They are generally referred to as a first order optimal strategy (FOOS) and are usually put in place intentionally.
Except you appear to have not read the link to see it doesn't apply here.
The example given is mage vs fighter. The mage is harder to use, but with better results for a skilled player. The fighter (the FOOS) is easier to use, but with
worse results for a skilled player.
Same applies to the second link's example of the "big money" strategy, which is simple, but is said to probably lose to experienced players. Or the spamming a single combo in fighting games, which I shouldn't even need to explain.
Both your links give the impression that FOOS is not really an unbalanced aspect. Instead, it's one that is balanced by skill requirement: low skill floor, but also low skill ceiling. It is not something that actually gives close to the best results possible in the hands of a skilled player, so it doesn't invalidate the need for skill.
But in Bannerlord, the Khan's Guard is simply all-around better. Whether you're skilled or not. So, it does invalidate the need for skill. It was pretty much all anyone used in Bannerlord Online from what I understand, until that mod nerfed the Glaives into lances.
In addition, the other stated benefit of the FOOS is to provide an easier option to get through the game, but Bannerlord already provides this with difficulty settings. You don't need a specific troop to be massively OP to make the game easier for newbies, they can just turn player and troop damage down to 25%, boost their speed, boost their recruitment, etc. Using faction specific troops as a difficulty setting is nonsensical.
In fact, if we want to make Bannerlord easier for new players the battles and troops aren't even really the hard part anyway (since the player can usually outrun parties they can't take on), the hard part is the horrible lategame game design that isn't intuitive or functional. So logically, one would start with that.
There is no good reason why the Khan's Guard needs to be so overpowered.
Because it doesn't allow a player to feel clever for having learned something about the game. Most players never reach those final perks, especially the Trade one.
(Although, honestly, Spring of Gold could just dump 25K in your lap per day and still not be particularly unbalancing. No one with Trade 250 or whatever is going to need money and anyone else will make about 30-60% of that just fighting lategame armies consistently.)
Whoever feels
clever from recruiting Khan's Guard and seeing them massacre every other troop in the game and saying "wow, I'm so smart for figuring out these troops are good"! would probably also equally enjoy someone jangling keys in their face. The answer is blatantly obvious.
If Khan's Guard were nerfed to a more reasonable state, there would still be "best" troops for certain situations, and players would still get the opportunity to work out which ones. If anything, it would be much more rewarding, as the answer would require some basic thought.
So: are you saying you actually support Khan's Guards being this broken? Or are you just arguing about quibbles for the sake of it?