But in my system of values, diversity and a multifaceted experience play a key role.
That is, first and foremost, I work on high diversity (this is what makes balance interesting), and secondly, I fine-tune the optimal balance to suit it, avoiding cutting back on diversity.
Moreover, I prioritize combat diversity above all else. I am willing to sacrifice interesting progression (leveling up 10 levels is not as interesting as 20), and I am willing to forgo the most powerful artifacts, such as +50% damage, just so that these factors do not harm the existing combat, where the abilities and parameters of units are fully revealed.
However, what I am not willing to do is reduce the role of individual parameters or abilities. That is, no reducing a 10% difference to 5% (parameters here play twice as strongly as in the original, and I see no point in making a smaller step). And no weakening abilities (except perhaps for BB), on the contrary, abilities need to be made more vivid and impactful; this always increases interest (better quality than quantity).
For example, 10/20/30% attack is interesting, while 5/10/15% is a pathetic remnant of a skill. On the other hand, with a large number of simultaneously taken skills for physical damage, 30% attack will exacerbate the gap, so the total number of skills in a build should not be large, that is, the hero's level should not be high.
Added 53 minutes later
That is, it turns out that:
The highest priority is high diversity, which creates a multifaceted experience and ensures high interest (hence the desire for strong and spectacular abilities, as well as "unconventional" types of units).
Balance and dynamics are secondary things that should serve this diversity.
A game with high dynamics but without gameplay is meaningless. As an example in the discussion, some maps in Heroes 3 were mentioned, where utopia is taken in 1 (!) week, due to the fact that a lot of wyverns/angels are farmed on snakes/canned creatures on the map, the same army is broken through many times per turn in a chain, and so on. It's fast, dynamic, but monotonous - there is zero interest in it.
A game with excellent balance but little diversity is dry and uninteresting. Chess is not the best example here; I would rather cite checkers as an example - a game where multi-step and deep calculation + combinations are important, but the basic diversity of pieces is very low, which is why interest in the game is quite niche. Getting a person to play checkers against modern games requires skill.