NHF – new heroes for Heroes V 3.1.

A decent, though somewhat one-sided, post.

I wrote practically the same thing, but not in as much detail :)

So, I provided a nearly opposite example for a binary skill system, which will result in significantly less randomness.

It's not a given; it needs to be calculated or verified. Because with a binary system, it's easier to take basic perks, but harder to take advanced ones; it depends on the specific scheme. In a ternary system, advanced perks are very often linear (without branching) and are taken automatically. In addition, most of the randomness comes from whether or not you get the desired school in the first place; the choice of a binary scheme has no effect on this factor.

It can be said with certainty that the downside of a binary system is its excessive rigidity, while a ternary system is more flexible in this regard (and the most flexible without connections, but that leads to wild randomness).

And resorting to personal attacks is not the way to go.

Personalities are irrelevant; the point of that phrase is that it is important to consider not only theory but also practice.

Added 16 minutes ago

Basic perks should enhance the entire school and, ideally, all skills afterward, because they are transitional. This is logical. However, the difficulty lies in devising such a system. If it were possible to program everything from scratch, there would be no problem. In other words, to create the word "ETERNITY" from the letters "O", "P", "Z", and "A".

Why should they? How is this concept better than others? This is unclear. There are almost no prerequisites for it; many attempts to connect certain perks as an enhancement are, to put it mildly, far-fetched.
There are always limitations; within the framework of a Heroes mod, they are one thing, and within the framework of a supposedly "free" game design, they are another. Optimal solutions must be derived based on the system of limitations, rather than complaining about them ;)

Take, for example, the classic combination of a tent and a plague tent. The first perk enhances the healing of the tent, allowing it to resurrect units. The second perk gives the tent the option to shoot. Neither the first nor the second perk enhances the other in any way. Shooting and resurrection are alternative actions. Yes, they enhance the tent itself, and the tent is the connecting link. But the fact is that these perfectly connected skills do not enhance each other, therefore they already do not fit your concept. What more can be said about other, less obvious connections.

In my opinion, the only limitation on connections should be that they should be easily associated, therefore memorable and understandable. And how the association takes place – by enhancing the same action, by a common object, by a common element, or even by name or extra-game meaning – is not so important. I also added a limitation in my mod: no branch should be overloaded (similar to how binary schemes are overloaded), i.e., there should not be fundamental perks that must be taken in order to take many others. Specifically, no perk should have more than 2-3 child perks (in my scheme, there are 3 child perks only in one school + ultimate perks in racial ones).