Problem is that "feel balanced" is a subjective thing. Playing on a horse and leading cavalry charges is going o feel different then playing a Robin Hood with band of archers and very different playing Viking raider in a shieldwall.
Sure, but I am simply pointing at the fact that demanding arbitrary buffs on particular units is not going to make anything balanced. How do you want to measure "balance" between the horseman and an archer for example? Should horseman be able to kill archer or should archer be able to kill horseman? Well the answer is both but then how exactly do you want to accomplish that? If you make horseman "stronger" so that he can kill archer then archer is not going to be able to kill the horseman and vice versa.
And the obvious thing I am pointing out is that horseman is not stronger then archer and archer is not stronger then horseman, there is no balance between them. The success of one or another is in how they are been used, not what "stats" or "buffs" they have.
Take a chess, which is simplified battle. What is stronger? Pawn or knight? Answer is none. Pawn can kill knight and knight can kill pawn. Trick is not in their "strength", trick is in the way you move them. Pawn doesn't have more or less HP armor or "attack" then a knight. If TW was able to make this aspect of combat more important then we might not have these discussions about buffing, depuffing and balancing this or that unit because it wouldn't matter as much.
Archers can attack at the distance. Cavalry can move and change position fast. Heavy infantry is good at holding position and fight in formation frontally. Make these things matter more in combat, meaning making combat more tactical and it would go long way towards "balance" without need to buff anything. Within limits of the AI and game system of course. I am not expecting complete realism.
There is not hard to find that T6 Archers will always be much more useful and appealing than T6 cavalry. Not sure if you have checked this video, but for me there is no reason because I should not get Fians over T6 cavalry 100% of the times (aside from me disliking to use exploits that make the game less enjoyable, in the same way I do not use smithing currently because it is broken):
I am pretty sure that the devs' tests are going in the same direction, mostly because it is pretty evident that you can defeat the AI easily if spamming archers (00:15:40 for archers battle), while cavalry is much harder to manage effectively and while the result is much worse. The balancing is pretty subjective, yes, but if you are getting infinitely better results if recruiting X unit instead of recruiting Y unit, there is not needed to be a balancing expert to know that something is really wrong with these results.
Maybe you want cavalry to be relegated to a 100% support role, while other people want cavalry to a more deadly role, etc. Who is right on this? Well, who knows, but what is pretty evident is that archers make feel cavalry and infantry useless because you do not need anything else if spam archers. If you get a different unit, you clearly know that your army is underperforming compared to an archers spam army.
Concerning the Archers vs Cavalry question, well, if the cavalry is able to reach archers in melee, cavalry should be able to inflict massive kills on these archers, but it is not happening currently because cavalry AI is mediocre as best and miss a lot of hits. Anyway, there is not much sense to argue about if cavalry AI should be improved or not, because at this point all people do agree with everything related to improving AI, plus devs have already confirmed that cavalry AI is getting improved, so people complaining about cavalry bad performance can be happy now.