Heroes of Might and Magic III
1 Million Gold 44 Week Production Battles
INFERNO VS DUNGEON

Score: 477,200

Comments:
- Inferno is so fun to use.
- We know the Pit Lords get targeted first. We put the Black Dragons in a position so that they’re aligned with the Scorpicores to get some 2-hexed damage from the Cerberi.
- Quick Efreet and Devils get to tie up Dungeon’s archers early
- The Black Dragons’ positioning in the 2nd turn was not planned, but it looks like the AI takes this spot to continue attacking the Pit Lords which opens them up for some collateral damage against their own Scorpicores
- Dungeon with two units that can petrify actually have a very good chance in terms of a probability tree analysis to get at least one or even two events to happen at a reasonable probability. So this take, we will be content with the Efreet being stalled for 3 turns by the Scorps
- More 2 hexed damage with the Cerberi in the 2nd turn, also before the Minotaurs get to attack.
- The 3rd turn we get some more unplanned Black Dragon collateral damage. However, after some time, you have a fairly good idea which units the AI target.
- All this and we have the Devils and about half the Efreet left. That’s how good Dungeon is.
Rules:
- 1M Gold is arbitrarily a nice round number, and works out to 44 weeks of production. Stronghold arrives at the lowest cost to purchase all troops at 44 weeks for 1,079,320, we round down to 1M gold and use this as the common denominator.
- The reality in this game is that you're often gold limited, which ultimately determines what troops you can afford, which compares towns more fairly.
- There are flaws that should be mentioned as with any sort testing, and really...this channel is just for fun.
A) S to XL maps matter. With everything from dwelling costs, building logistics, to long term tactics like Rampart's treasury, Necropolis' Necromancy, and the Pit Lord's Demon Harvesting.
- Primary Skills: ATK and DEF are set to zero for both heroes
- Secondary Skills: No heroes have any secondary skills that affect damage calculation (Offense, Armorer, etc)
- Spells: No spell casting from heroes, except for Remove Obstacle to reduce luck factor on the terrain. Creature spell casting is allowed. Tower battles and Master Genies? in this case, I haven't come up with any fair rules to change anything. This is because with 3 Master Genie stacks, there is high probability in having at least one spell that can significantly alter the outcome of the battle. All battles that involve Tower will be highly variable.
- Morale: The attacking hero is equipped with the Spirit of Oppression artifact to negate all morale bonus on the battlefield to eliminate the luck factor of morale. This puts Castle and Necropolis at slight disadvantages - can't be perfect. The disadvantages are small.
- Terrain: Sand Terrain to negate terrain speed advantages for all towns.
- Scoring: Score after each battle is determined by the sum of all units remaining after battle, multiplied by the cost per unit. Resource cost is not included as the game isn't typically resource limited.
- Pre-emptive: We assume the attacker, and the opponent is defender for simplicity, like White player in Chess. The opponent is typically set up with units in slots that should make sense with archers in the corners and slower units protecting archers. And our freedom to move our pieces around as the attacker (also arguably with archers in the corner etc) reduces variation in these battles.
- Miscellaneous Luck: There is luck associated with creature skills that can significantly change the battle. These are allowed. However, with accuracy and fairness in mind, I choose battles that are reflective of fair probability outcomes. The Ghost Dragon's Aging skill is the best example here. The probability of Aging an opponent stack is 20% on attack. If we assume three stacks of Ghost Dragons (two stacks with 1 Ghost Dragon), Necropolis will have much higher than 20% to inflict Aging just once in just the first turn. In general, a fair way to approach this is to evaluate how many major turns occur on average in these battles, which often comes out to about 5 turns. This involves a probability tree analysis with successes and failures. This means that there is a 41% chance to inflict it once in battle, 20% chance to inflict it twice, etc etc. Only 41% chance matters here - this means for these videos, I will accept battles where Aging is inflicted once in battle, but will discard the battles where it's inflicted twice or more.

свернуть