A group of giants means they're guarding something tasty; you'll scrape together 2500 for an extra scout hero one way or another to check if it's there or not. But then having to report to your opponent how many times you did it and whether you broke the rules isn't the most pleasant thing.
The other extreme is deliberately leaving a weak stack of neutrals somewhere to bash, but they have a Growth Stack, and you used them several times — so what now, should I techport?
By the way, the probability seems to be less than 0.5 — people once researched the presence of Growth Stacks on an XL map, never found a pattern, but the matrix was always sparse enough.
There are games of chance (roulette), there are skill-based games (chess), and then there are semi-gambling games (poker).
Heroes in its modern incarnation falls into the semi-gambling category — that is, randomness significantly alters the course and outcome of a game, but over a long distance, skill still makes a difference.
There's just one problem — in poker, when you lose, you give up a pot whose size you control, and you spend at most a few minutes on a hand.
But in Heroes, with an unlucky morale proc, you can lose a game you spent several hours on. If Heroes games lasted half an hour to an hour — no problem, but as it is, it's a shame. After all, I'm a busy person and I don't have the opportunity to churn out games like hotcakes, hoping to overcome variance through sheer volume.
Moreover, if you constantly play around every morale proc and stack, your tempo will drop through the floor, which risks losing to a bolder opponent. In turn, he risks losing due to his own boldness in some battle before even reaching you.
This approach doesn't mesh well with the idea of determining the strongest player. Because ideally, competition should be directly against the opponent, not against the map. That is, the opening phase (1-2 weeks) is run on autopilot, a kind of mandatory program, and only then does the constructive action begin.
Alas, not everyone shares this view; many people like to flip a coin, run around neutrals, and then (if they haven't lost) compare in the Field of Fame what God has sent them.
Out of curiosity, you could compare the seasonal reports on HeroesWorld, HeroesPortal, and "Bay" — in the first two, a huge percentage of reports are like "117, game didn't work out" or "115, opponent lost to morale" or "at 216, showed the opponent a pack of Archangels, he surrendered." I used to glance at those threads, but in recent months I don't look there — at most one interesting post per couple of pages. But on "Bay," I regularly read the seasonal thread, and there are often interesting reports.
Or, from personal experience — I play quite often with TERM, both in Full and in Light. In Full, almost all games we played were long and meaningful. Regardless of template, settings, etc. But in Light, out of 6 games, 2 were conceded in the first week, and in another 2, we took auto-resolve due to completely absurd morale losses (like 4 stacks out of 5 getting morale in the first round). Who will be hit by such a distribution next is unknown, and there are also imbalances in starting armies. And this is on mm6; on a "sm" (simultaneous) map, with players of roughly equal and high skill, the swings are probably even wilder.
I often hear whining like "the first week in TE is a bit boring." Of course, it should be a bit boring; it's like in StarCraft/chess — the build order/opening — N minutes/moves where you just develop your economy/pieces and build troops, and only then does the meaningful part begin — battles with the opponent.
Right now, the technology for creating a fully symmetrical random (including the sizes of piles, banks, and artifact composition in Dragon Utopias) is ready. For now, I'm playing with people, testing it, looking for bugs.
This is a continuation of implementing the approach described above — randomness should strongly influence the course of the game but have little impact on its outcome. In this regard, mirrored random combines the best aspects of fixed maps and random maps — a random map, but with almost equal chances for the players.
Of course, in this technology, there is no place for morale, Growth Stack, significant variance in starting troops, or random level 2 dwelling construction. The only asymmetry that can't be eliminated is heroes in the tavern and skill trees. Oh well.
For these same reasons, there's no point in implementing it in Light — why, when its paradigm inherently contradicts the concept of equal starting positions?
