MMH5.5 - Heroes of Might and Magic 5.5

Detailed breakdown of classes (regarding differences):

Warriors, general template: attack + defense = 75-85%, spellpower + knowledge = 15-25%
- Offense
- Defense
- Luck
- Leadership
- Logistics
- War Machines
- Counterattack (Combat)
- Light / Dark
- Racial

(rare skills template: Education + 2 dampening)

Standard classes:
Ranger: Light 45/35/10/10 = attacking Light
Knight: Light + Dark dampening (instead of racial) 35/45/10/10 = defensive Light, no racial
Overlord: Dark + Light dampening (instead of racial) 45/30/10/15 = attacking Dark, no racial
Demon Lord: Dark 50/30/10/10 = attacking Dark
Death Knight: Dark 30/45/10/15 = defensive Dark

Interesting classes:
Engineer: Chaos dampening (instead of magic) 40/40/5/15 = neutral without magic
Chieftain: Education (instead of luck), War Cries (instead of magic) 55/30/5/10 = attacking cries

Conclusions:
1) almost all warriors are built according to one template
2) the differences lie in:
- a single school of magic - light or dark (or cries, or no magic at all - with dampening)
- an attacking or defensive distribution of parameters
- a racial skill (or none at all - with dampening)
3) Chieftain is interesting due to having education instead of luck
4) warriors can have either Light or Dark (but not both together)
5) Sorcery, Occultism, Chaos, and Summoning are unavailable for all warriors
6) Academy has no warriors

What I didn't like about the warriors: a single school of magic (no variety in magic) and the templated nature.
How it can be fixed:
- give several classes both Light and Dark simultaneously;
- give several classes accessible (not rare 4%) education;
- for several classes, move 1-2 standard skills from available to rare (4%), for example: DemonLord has rare leadership, Death Knight has rare luck, etc.

I will break down the other 2 types of classes a bit later. Preliminarily, I can say that they suffer less from templating and are more interesting.

Added after 34 minutes

Don't you think this kind of difference is worth 4 points of investment? What other combination of skills + perks can give such advantage at level 4?

I never claimed that active abilities in Combat are bad (and by the way, I liked the changes regarding hero attack damage). The point was that the Combat skill itself (without abilities) is very weak. Because of this, it's not profitable to spend 2 levels on Combat (Advanced + perk) for a small +1/+1 ability.

Added after 24 minutes

Balance was achieved by reducing Might heroes and improving magic ones.
The reason was might was reduced is because the pace of the game became too fast. Here is the math:
- with 5% damage boost per primary skill if the difference between unit attack and enemy units defense is 30, this is 30*5 = 150% bonus damage.
This means 1 unit hits as hard as 2.5 units without a hero. Here I do not even add Offense, luck, leadership and other bonuses which can make this difference as high as 4 times.
In a scenario where both heroes have high attack and low defense the winer is the first army to hit/shoot because it will at least cut in half or one shot any of the opponent stacks.
Where is the strategy here?

I wasn't criticizing the 3.33% parameters at all. That in itself is normal, specifically for long games.
The point was about small abilities. If in the original a good ability gives +2-3 parameters, then when converting to new characteristics, we should get +3-5 so that the abilities don't look worse. But in your mod, only +2 per ability is considered standard.
In short, in my opinion:
- parameters were weakened (from 5 to 3.33%) - this is OK (when calculating for a long game)
- abilities were weakened - this is not OK (since the "flavor" is lost)

Now lets discuss Might vs Magic heroes. In this situation Magic heroes will be on the losing side when it comes to frontal confrontation so they need to have their magic sharp.
But even if they have strong magic they need time to cast it. Magic heroes total damage output is equal to the sum of damage made in all rounds. The number of rounds depends on the survivability of his army. It does not matter how strong spells does the mage has unless he can cast for more than 3-4 rounds.

If magic is strong enough that 1 cast = minus stack (1 spell = 1 dead stack) (with sufficiently high spellpower), then these 3-4 rounds might be enough for a mage to turn the tide of battle (under sorcery, this is 5 dead squads out of 7 solely due to the hero).

Added after 41 minutes
Also, I'll give a general tip for your 3.33% characteristics.
At the start of the game, +4 to offense (for example) from an ability won't break the game. But +4 to spellpower might (because you didn't weaken spellpower, as far as I know).
Therefore, I would tie bonuses to spellpower and knowledge to levels, but never bonuses to offense and defense.

Examples of abilities giving bonuses:
1) +4 to offense - OK (a good, tasty ability, we'll level it)
2) +1 to offense for every 5 levels - definitely not OK (we want a tangible bonus immediately, not a measly +3% damage)
3) +4 to spellpower - possibly not OK (OP in the early game?)
4) +1 to spellpower for every 5 levels - OK (even +1 to spellpower is noticeable in the early game)

Added after 47 minutes
A couple more remarks appeared regarding the undead:

1) Retribution: for Necropolis heroes +2 to attack
2) Leadership: for Necropolis heroes +2/+4/+6 to attack

And what if I take an undead hero with living troops (another hero's)? Will I get additional bonuses just like that? From expert leadership +3 morale and +6 attack, and from retribution - +10% damage and +2 attack?

This is one of the reasons why I suggested moving retribution into leadership. And accordingly, exclude leadership from the progression for undead classes, including Death Knight.

Added after 7 minutes
By the way, regarding herds. If I take a Knight for a dwarf army and level up his herds, will he get his +2 to offense or defense in addition to the native effect of herds?

That would be strange.

Added after 41 minutes
Continuation of class breakdown, Balanced type:

General template: attack+defense = 50%, spellpower+knowledge = 50%
- Defense
- Luck
- Logistics
- Education
- 2 schools of magic

(Light, Summoning) Paladin = Leadership, Counterattack, dark dampening 15/35/20/30 = defensive Light+Summoning
(Light, Summoning) Warden = Paladin 15/35/25/25

(Chaos, Dark) Gatekeeper = Offense, Sorcery, racial 35/15/25/25 = attacking Chaos+Dark
(Chaos, Dark) Assassin = Offense, Sorcery, light dampening 40/10/30/20 = attacking Chaos+Dark

(Summoning, Dark) Reaver = Offense, racial, light dampening 35/15/20/30 = attacking Summoning+Dark
(Light, Summoning) Seer = Offense, racial, machines 30/20/20/30 = attacking Light+Summoning
(Chaos, Light) Runemage = Machines, Sorcery, racial 15/35/30/20 = defensive Light+Chaos

Strange class:
Shaman = Offense, Machines, Sorcery, Occultism, chaos dampening 15/25/35/25 (exception to the 40% on 60% rule)

Conclusions:
1) Paladin = Warden (almost completely)
2) Gatekeeper = Assassin (difference in racial and its absence)
3) other classes are better developed (no obvious repetitions)
4) no one has occultism (except the strange Shaman)

What pleases me: 2 schools of magic, less templated (compared to Might).
What doesn't please me:
- 2 pairs of duplicating classes (Paladin = Warden, Gatekeeper = Assassin);
- too even a distribution of parameters, without clearly expressed traits

Added after 54 minutes
Final part, breakdown of Magic classes:

General template: attack+defense = 20-30%, spellpower 40-50%, knowledge 25-35%
- Logistics
- Luck
- Education
- Sorcery
- Occultism
- Chaos (Destructive)
- Summoning
- Light / Dark

Sorcerer: Dark, racial 10/10/45/35 = dark caster
Necromancer: Dark, racial 10/15/45/30 = dark caster
Warlock: Dark, Offense 15/10/50/25 = dark caster without racial

Druid: Light, Defense 10/20/40/30 = light caster with defense
Flamekeeper: Light, racial 15/10/45/30 = light caster

Elementalist: Light, Dark 10/10/50/30 = archmage

Special:
Wizard: Light, racial, Defense (instead of occultism) = 10/15/30/45 (non-template parameters)
Heretic: Dark, Offense, Light dampening (instead of Summoning) = 30/15/35/20 (non-template parameters, do not correspond to Balanced)
Witch: Offense and Defense (instead of Chaos and Light/Dark), Necromancy = 20/10/40/30 (only 1 school of magic - Summoning)

Conclusions:
1) almost all mages are built according to one template
2) the differences lie in:
- which school will be the third alongside the Chaos+Summoning pair - Light or Dark (Elementalist has both)
- racial (or its absence)
3) but there are also unique classes that break the template (Wizard, Heretic, Witch)
4) logistics and luck are a universal pair of skills present in all Might, Magic, and Balanced classes
5) leadership and machines are absent from all Magic classes

What I didn't like: templates again; necromancy for living heroes; eternally high spellpower 40-50%.
How to fix:
1) put some classes on alternative magic trios: Light + Dark + Chaos/Summoning
2) remove education from some classes
3) remove occultism from some classes
4) put some classes on leadership (empathy)
5) remove necromancy from living heroes
6) balance the game so that classes with high knowledge (similar to Wizard) are viable and introduce corresponding classes

Added after 3 minutes
In conclusion: from the point of view of uniqueness, Balanced type classes were the most successful; they are not as one-dimensional as Might or Magic. However, there are overlapping classes there too.