Caster of Magic II Bug Reports!

[Intaka]

Try surveying map location 2, 35. I think I have it nailed down repeatably. It appears the world map tiles around Freyas razed capital are causing the crash.

Same game. Pushing forward to try to help narrow this down. Had an unrepeatable fatal error on processing of End Turn. March 1514 -> April 1514.

So there's some corruption somewhere. Did you have any luck with the 2, 35 coordinates?

[Seravy]

Yes, 2,35 crashed even without clicking anything.
It's crashing because the X looping function call is missing from the TileMaxPop function so after 1 it goes to 0, not 44. I'll upload the fixed version ASAP.

[Seravy]

2020-12-15
-Fixed bug : Scroll buttons for enchantment list on outposts is at wrong location.
-River tiles now cost 1 in combat for water walking units
-Fixed bug : Error message when closing the game while the save thread is still processing.
-Fixed bug in adjancency formula between planes
-Changed adjacency formula to be less strict
-Fixed bug : Surveyor crashes at X coordinates near the edge of the map.

[Intaka]

Another bug down. Thanks for working with me on that one. It wasn't the obvious kind.

[Xanatar]

Hello I don't know if it was already reported or corrected but when a unit participate a combat while under the "city wall bonus" it is keps in world map when looking at the unit. Still it is dispelled when another fight start so it is not really a problem.

[jhsidi]

Is this an example of your wizard destruction mechanic at work, Seravy? Horus, fortress at 50x24x1, gets wiped out by Lo Pan's stack, at 53x24. https://ufile.io/h2tr3251

Horus: 3 heroes, 6 magicians, Heavenly Light, plenty of mana
Lo Pan: 6 wolf riders, 2 hell hounds

Seems a bit overpowered. That wolf rider stack would get thrashed in an actual fortress battle against magicians. It's a bit bad for me as a player because it's also unpredictable -- if I thought Horus had a chance of getting wiped by that stack then I would have eked out any benefit from I could get from him before he died (such as trading), or maybe even tried to wipe him myself first.

In general I think that even after your recent changes, there's still not enough war / too many peace treaties. Granted, in this save there are several wars going on. But it's 1509, and a couple years ago, the world was mostly peaceful. Moreover -- the wars generally don't seem to accomplish enough, with power rankings not changing a great deal over the course of a war. I especially wish my own wars had a more visible effect on the world.

For example, Oberic was at war with me until a few turns ago. I took several of his cities and his power rankings dropped lower very quickly. However -- he didn't lose any of his many wizard pacts, and nobody else declared war on him, even though I have multiple allies. Similarly, I would have liked to see some of his allies threaten me or full on go to war with me.

Have you tried out these things yet?

• Incurring diplomatic penalties against wizards who rapidly lose overall power
  
• Amping up the diplomatic penalties for war against allies
  
• Maybe increasing the negative relations score for opposing magic types by 1.25x or 1.5x?
  

I'd be willing to play few test games under those conditions.

[Seravy]

Save file isn't compatible, it seems to be an older version.
I can't load it.
Check the log file, it should say something like "Horus is an elimination target" if it's that mechanic.

There is definitely more than enough war - The AI can already barely afford the mana cost of battles as far as I can tell although of course rolling 13 peaceful wizards are a different story but in average cases.

Wars don't accomplish anything for the exact same reason why the elimination mechanic was needed.
The AI combat simulation is too accurate. There is no way for the AI to ever beat another AI because a stack of defensive normal units, usually magicians, will almost always beat a stack of offensive normal units, like wolf riders. Which kinda has to be that way otherwise human players couldn't hold their cities either and the game would be unplayable.
This isn't a problem when the human player plays : they can intentionally build a stack to counter the magicians, like cast Elemental armor or even resist elements on those specific 9 units that will fight. The AI cannot do that because it has no brain and no planning capacity. If it has a stack that has those buffs then yes, it'll notice that stack can beat those magicians and attack. But it cannot specifically build that stack intentionally. It first has to pick the Elemental Armor to cast based on their Objective, other spells available and random chance. Then the spell will always target the unit with the highest buff priority which in general will be the strongest unit in the overall strongest stack. While this helps accumulating buffs on one stack, it takes a while and during that time the stack will probably attack something else and/or regroup and swap some units. Even if the stack somehow does get buffed, if it's not already on the continent with the city of Magicians, then it needs to take a boat, and get shipped to somewhere with targets which might or might not be a continent that has a city full of magicians. And even if the AI manages to get through all that successfully, they'll notice there is another city on the continent which has 2 more buildings and population, so it's a higher value target, and go attack that one then leave the buffed units garrisoning there.
Finally if the stack does conquer the magician city, it'll stay there to garrison it unless it was an exceptional case of being an intercontinental summoned doomstack, so the other two dozens of cities that have magicians in them are safe for another 20-30 turns until the AI manages to build another such stack.
Ultimately, the AI will have very few stacks capable of cracking the average garrison strength of the same tier AI or even the tier below by one. Pretty much only the stacks of 9 summons will be able to do that. And even those will fail on something like a fortress or a Guardian wizard.
There are of course cases where the AI has better chances, for example if they are a Chaos wizard. Magicians have tiny hit points and defenses, so a Chaos wizard might be able to outright destroy the entire garrison solely by having enough casting skill, good enough direct damage spells, and an army durable enough to survive those 3-4 turns needed to destroy the magicians. If that's what is happening, again, you can tell from the log file. It sounds unlikely though because Horus is probably life/sorcery which basically means they counter direct damage by buffing army hit points/defense.

Anyway, what I'm trying to say is it's not a diplomacy problem. It's the direct consequence of how the AI is required to play. They must prioritize garrisoning high because the AI leaving cities poorly defended would be seen as a bug by players. But human players are smarter at attacking cities so any garrison built to stop or stall a human player will be even more effective at stalling the other AIs unless they are literally two tiers stronger than the other AI but for that advantage to develop the game has to be going for like 12-18 years depending on difficulty levels.
In addition, even if that wasn't an issue, an AI war can only be effective at conquering if all of these are true :
-The AI has a LOT of unused, quality troops
-They have enough naval dominance to not lose the troops on the way to the target
-They have enough mana crystals to actually support those armies in battle
-Those troops don't get killed by a different AI who is even stronger
-The human player doesn't interfere

About the diplomacy part though, AI wars, like human player wars, are more likely to end with a peace treaty if there are no fights going on. So as soon as an invasion is stalled because the cities are too well-garrisoned, and the AI changes their target continent and removes the troops, it'll eventually lead to peace, if it was a war between near equal powers. This is not a problem though, obviously if the AI can't conquer anything anyway then ending the war works better for them.

[jhsidi]

Yeah I'm, I think, two versions back. I can't play every day so in order to have a continuous game I didn't update for a while.

I understand what you're saying -- most strategy games have the problem of AIs that can't beat each other, to some degree. The "marked for elimination" tactic is valid in that sense. However.. it still looks and feels pretty off, because the bonus is massive and only applied in specific situations.

One alternate thing you could do is flip the mechanic around to silently provide some bonuses to AIs that have been selected for advancement. Call it the "marked for greatness" mechanic. Give a bonus to income and a small strength bonus to all fights with other AIs. That way they could grind their way toward victory and it would be less jarring than seeing a stack of trash conquer a capital full of elites. I think this advantaged AI should also refuse peace treaties, since quite often I've observed that AIs won't stay at war even if they're winning.

I can think of a few strategy games off the top of my head that work that way, providing bonuses to favored factions and removing bonuses from unfavored.

[Seravy]

Well, the bonus mechanic and the elimination mechanic would be pretty much the same thing, really.
They are the two sides of the coin.
The feature triggers when a wizard not chosen for elimination fights one that was chosen.

Giving a bonus to the not chosen one instead of a penalty to the chosen one would lead to the same result.

A better solution would be to disable the penalty if the human player has units within a certain range from the battle.
There are two problems with that though.

1. The human player will usually leave units near a leftover enemy fortress they can conquer, to kill units that come out of it. So in the exact case where the game mechanic is most needed, it can't trigger.
2. The human player will remember what was in the fortress so even if they left, if they see the fortress fall, go back and see it got conquered by some wolf riders, it's no different from it happening when they see it directly.

"Grinding towards victory" would also not solve our problems. It's too slow. When I ran the tests in the older versions, the wizards stuck on having only 1 city left generally managed to survive another 10 years before actually getting beaten. So no amount of economy advantage was enough to make it happen before researching very rare spells.

[Anskiy]

I don't see much of an issue in the situation Seravy described. Stacks of magicians are a formidable defensive force, and elemental armor is not a cheap counter to them. It costs a lot to spam it on nine good units, and it does nothing to protect against armor ignoring or resistance based effects, of which there are plenty in the game. If the game has got to the point where cities are filled with stacks of them, it should require better spells or units to defeat it. However I do think teaching counter buffs (and counters in general) to the AI would be a good idea.