Caster of Magic II Bug Reports!

[jhsidi]

Even if this was real normal combat the cockatrices and berserker would kill the demons by turn 3. The spells from the demons will most likely do damage but ~25 is a realistic estimate (the globals don't do damage except gate of hades which doesn't do much in a short battle either. Direct damage spells do but each demon can only cast one and half of them will cast globals instead. So it's like 3-4 Reaper Slashes and that's all the damage the demons can do.)

I strongly disagree -- you should try it. Every time I take a stack like this in to fight demons, it's a massacre.

The casting of black prayer and terror would prevent most of the units from even moving. The berserkers, with their low resistance, would have a 0% chance of moving. The cockatrices and sprites would stay frozen long enough for the gate of hades to eat away at them. A total of 17 damage may be reasonable if the demons had a weak ranged attack, but it's a very poor representation of what would happen in a battle. Even if the RNG caused the demons to solely cast reaper slash they'd at least take out the sprites and hero.

In general, I feel like spell casting units are undervalued. For instance, a stack of angels and arch angels is one of the most typical tower defenses and because of their values, I'll often see stuff like 5 angels + 3 arch angels. And those are the hardest towers in the game. Stuff like death knights, wraiths, stone giants, doom bats etc simply does not compare.

Anyway, one more possible bug for you: I just made peace with Lo Pan this turn. Over the end turn he (sometimes) declares war again. Is that because he's maniacal? Or is a countdown timer not working? https://ufile.io/he73p354

[Seravy]

The casting of black prayer and terror would prevent most of the units from even moving. The berserkers, with their low resistance, would have a 0% chance of moving. The cockatrices and sprites would stay frozen long enough for the gate of hades to eat away at them. A total of 17 damage may be reasonable if the demons had a weak ranged attack, but it's a very poor representation of what would happen in a battle. Even if the RNG caused the demons to solely cast reaper slash they'd at least take out the sprites and hero.

Yes but that's a "hit or miss" situational thing that can't be considered. If you have Resist Magic or Bless known, Terror has pretty much no effect on you (you just need to make sure you cast it on the 2-3 units that can kill the demons first.)
Also, unless the entire army can't hit fliers, the demons will come and attack you anyway, killing themselves.
Anyway, it's impossible to convert situational spell effects into an accurate result because they are nonlinear. Have 39 MP and you can't cast Terror or Gate of Hades, but at 40 you can. So 1 MP does all the damage? Obviously not.

Even if the RNG caused the demons to solely cast reaper slash they'd at least take out the sprites and hero.

Except, that is the same thing. 3 Sprites + hero = ~25 damage. The numbers match your expectation.

Damage distribution rules AFTER the battle are a separate matter. The calculated result is literally "Attacking army wins and takes 25 total damage" - which can mean either the 3 sprites+hero dead or everything alive but with 3 damage on each unit - the damage distribution algorhitm decides that after combat.

By the way, combat calculation itself contains no random element currently, which means it's always the same result - same total damage to be distributed. This also makes it reliable - if the AI simulates the battle and the result is "we win" then they will win. There is no room for bad luck. The AI condition to attack is "Yes if expected damage is no more than 50% of the army HP, or 30% if we have a hero.", so in worst case they can only ever lose half of a stack but considering the damage distribution rules, it's much more likely they only lose 1-2 units and have most others slightly wounded. If necessary we can tweak these numbers to make the AI take slightly more risks in exchange for becoming more aggressive and able to clear lairs earlier, but they will lose more units in the process. (and of course their chance to win in real combat vs the human player will be worse if the change applies there as well.)
30% and 50% might be overly cautious as with those numbers the AI pretty much never loses more than a few weak units from a stack.
In CoM I the AI needed to be something like 25% stronger than the target but that's hard to convert to this new metric. The old system also had random outcome so it was more like "high chance to win, very low chance to lose and medium chance for some serious damage to the stack" while this 50% if literally 100% chance to win, 0% chance for taking serious damage.

So to make it short, the only random element is, how the final damage is distributed between the units in the surviving army which generally tries to minimize lost units for the AI player while maximizing it for the neutrals - as that is how it most often works when the human player fights as well. Players will focus fire on the neutrals killing as many as they can, but will also position units properly to make sure nothing dies on their side whenever able. Obviously, simulating the "whenever able" part is impossible nor is it intended - the AI losing more units doesn't make the game better in any way.

[Seravy]

Anyway, one more possible bug for you: I just made peace with Lo Pan this turn. Over the end turn he (sometimes) declares war again. Is that because he's maniacal? Or is a countdown timer not working? https://ufile.io/he73p354

Bug, the peace counter was literally set to the amount of relation instead of the peace counter turns. Will be fixed in the next update.

[jhsidi]

Something is odd about neutral air elementals. I've been noticing that when neutral monsters attack me, I can't seem to scout where the air elementals are even when they should definitely be in a specific place. For example, if on turn 2 I know they'll attack my bowmen in the back of a city, then on turn 1 they have to be somewhere near the walls on the right or left of the city; but checking these places doesn't find anything.

Also, they seem to disappear each turn after attacking. I don't know if that's an AI routine for them to hide after making an attack, but it's pretty buggy looking as is.

Check out this save file. Take the active stack into the adjacent layer at 48x20 and watch closely... on the first turn, the air elementals are briefly shown moving away from the great lizards. That's impossible; they only have enough move to reach the lizards. They're actually about 3 tiles back. https://ufile.io/z2k76tde

[Seravy]

Check out this save file. Take the active stack into the adjacent layer at 48x20 and watch closely... on the first turn, the air elementals are briefly shown moving away from the great lizards. That's impossible; they only have enough move to reach the lizards. They're actually about 3 tiles back. https://ufile.io/z2k76tde

that's actually 19x48 btw.

Definitely looks abnormal.
...surprisingly, after giving True Sight to one of the lizards, the elementals moved normally.
Yeah, the invisible units were set to move fast enough that the gametick when they set their position is skipped.

[jhsidi]

Dispel Magic is a bit messed up. You might have to test this on your own since enemy AI does it in battle unpredictably but:
1) An exception is always thrown when an enemy wizard casts dispel magic. The dispel animation always plays in the wrong location and I've never seen it actually dispel anything.
2) When I used dispel magic on a confused unit, the confusion was removed but the unit still had the confused effect. After battle it survived but during the battle I wasn't able to control it -- it also retains the confused effect in later battles.

You can try using this file for testing #2, my great lizards get attacked over the endturn and the enemy may cast confusion: https://ufile.io/df7563ez

[jhsidi]

Bugs:
- AI wizards can continue signing treaties with other AIs while banished
- Casting Elemental Armor on the overland map has no result; casting it in battle causes an exception
- Purify is kind of hard to use because it has to be renewed each turn, I think it may be a bug that both build road and purify don't work continuously as they did in MoM.
- On "All units retreat exhausted" results I'm losing 1 fame sometimes. Usually it's when I kill all of their units but one invisible one, and I'm not taking any losses. Maybe the opponent is supposed to be losing the fame?

Notes:
- What is poison blocked by? I have units with both resist magic and resist elements on them but it seems like neither one does anything against poison.

[Seravy]

- What is poison blocked by? I have units with both resist magic and resist elements on them but it seems like neither one does anything against poison.

Absolutely nothing at all besides raw +X resistance stats.

[Seravy]

- Purify is kind of hard to use because it has to be renewed each turn, I think it may be a bug that both build road and purify don't work continuously as they did in MoM.

Couldn't reproduce this one. I set a shaman to purify and it kept working until the corruption disappeared from the tile.

[Seravy]

- On "All units retreat exhausted" results I'm losing 1 fame sometimes. Usually it's when I kill all of their units but one invisible one, and I'm not taking any losses. Maybe the opponent is supposed to be losing the fame?

Did you summon anything? Combat summons do count as lost units unless we change the rule, so if you aren't winning the battle,and you summoned at least 12 experience worth of creatures, you lose 1 fame.