Possible changes for the next update

[Seravy]

I'm considering adjusting the following.

1. Building destruction rate when conquering.
This was originally added to balance out the resource advantage the AI gains though difficulty modifiers, so the player can't take advantage of the AI building significantly faster, and as a secondary effect also to make conquest slightly less rewarding compared to other ways to invest your resources (lair hunting and building up city economy).
While the latter is still valid, the AI resource advantage is much lower nowadays so I wonder if maybe these percentages could use some adjustment. (In particular the rate on Normal, Fair, Advanced and Expert seems somewhat questionable and exceeds the AI's production and population growth bonus by a significant quantity)
As a potential side effect, AI players conquering each other also comes with less destruction, which might in theory lead to stronger endgame AI if they had wars and conquered many cities. So this might not actually improve the player's chance at winning as much as it sounds like but might make the game more enjoyable by conquering more relevant high value buildings in wars.

Should I lower the destruction rates or not?
Additionally, the Smithy has double the normal chance of getting destroyed, maybe this should not be a multiplier but a flat addition instead (30% higher chance seems right, if we lowered the base values)

2. Chance of hiring heroes.
I think this is fine for hero 1, 2, 5 and 6 but maybe the third and fourth hero could have a slightly higher chance than 1 in 3 and 4 years respectively?

3. Power Link
Is the AI's reaction good as is, or would it be better if I added a percentage chance for the AI to skip casting spells for a turn if it rolled a low cost spell and has no rare or better summons available?
Or is the wizard elimination mechanic and the power cap on the spell already good enough to balance it? (assuming the player didn't disable the "no early very rare spell treasure" feature)

4. Spells found too early in treasure
I do believe the 1 book or retort pick given when a Very Rare spell is replaced is fair but I'm somewhat uncertain whether the 1600 value item is really good enough compensation for the player for missing out on a rare spell?
I thought it was the item plus an uncommon spell but apparently not.

[Anskiy]

1. I do think that the building destruction rate could be lowered a tad, as it stands it's too common to capture cities that are nearly worthless to the player, making wars feel like a continuous grind on resources. On that note, how about bringing back the instant razes as an option, perhaps through modding.ini.

2. Don't have any strong opinions on this.

3. I think the wizard elimination mechanic could stand to be a bit more aggressive, it's still common for a whole bunch of them to survive even by 1520 or 1525. I think it doesn't kick in when the player is considered the "weakest" wizard, which is a flaw. The player's current state should have no bearing on this, after all having fewer but more dominant wizards is more balanced than having many but weaker wizards, and the further the game goes the more it should tend towards the former.

4. Don't have any strong opinions on this, but I do think some of the high tier rewards could stand to be made more exciting. Finding gold and mana from lairs containing very rare or even rare monsters feels a little disheartening.

[Seravy]

I think the wizard elimination mechanic could stand to be a bit more aggressive, it's still common for a whole bunch of them to survive even by 1520 or 1525. I think it doesn't kick in when the player is considered the "weakest" wizard, which is a flaw.

I checked the coding to verify this.
The conditions to mark an AI for elimination are the following :
-Historian graph total cannot be more than 75% of the average among all players
-Historian graph army cannot be more than 75% of the average among all players
-There cannot be more "weaker" wizards than this wizard than the among that would bring the total player count to the "desired quantity" calculated based on the settings in the INI file and turn count.

The human player is not counted in the number of weaker wizards, but is counted in the average historian levels.
So if the human player is extremely weak, they might bump the average lower which makes it harder for any AI to meet the 75% requirement but probably this effect isn't significant.

(with 8 players, a human player with 0 on the charts would drop the average by only 12.5%)

However, I do think the human player having any effect on this is probably not optimal, and shouldn't delay AI retirement even by this minimal quantity. Albeit logically, the AI player(s) seeing someone weaker than them should make them want to stay in the game longer, even if that person is the human player. It's also generally not ideal for gameplay when the human player is already weakest and the weak AIs they have a chance against leave the game. Although if they notice what happens and gang up on the weak AI while the other AIs are attacking, it can help the player a lot, too.

This is merely the initial requirement for "elimination" to trigger, and it takes a very long time before the AI actually leaves the game. In particular, it only makes other AIs declare war. Another 3 years has to elapse and the targeted AI must lose enough cities to have only 1-5 of them left depending on the map size, before they retire for real.

The real problem might be the lack of AI wizards who clear both "75%" requirements early enough. If none of the AIs fall behind significantly enough... but in this case, the human player dragging down the average by 5 or 10% might make an impact afterall.
Let's say instead of 75% it is 66% due to the influence of the human player. That's relative to the average so it's two thirds of strength compared to the middle player and probably half or less than the strongest AIs. And this has to happen simultaneously on both charts, so a weak player spamming Berserkers or Gnolls would likely avoid it through the military chart, while a player with no military but decent economy might avoid it though their power income and population pushing it up.

I think I'll remove the human player from the pool when calculating the average. I'll probably also lower the feature to work at 7 wizards or more, instead of 8.

[Desertfox]

I skip most offers of mercenaries because they typically aren't cheaper than building units myself if I use the money to force buy.
Perhaps the initial hiring price of mercs can be lowered?

[Slingers]

After trying things with More Stuff MOD - Advanced/Expert difficulty:

1.) I used BuildingDestructionRateCap = 50.

2.) Heroes I gave Fame requirement to hire. I got the impression it was allright.

3.) Elimination mechanic I set back to default, as it seemed to be on point.

4.) One tier lower spell + item seems ok. Do not know if it works.

[Seravy]

One more problem I found in the elimination system.

If N wizards should be chosen for elimination, only the weakest N wizards are valid to be chosen according to military strength.
However, they have to be below 75% of the average on both military and overall.
What this means is, any wizard that's in the group of "weakest" who has over 75% average total historian power, blocks the elimination of a wizard, as it occupies the "slot" but can't be chosen. This might be temporal if they eventually get behind in other statistics too, but can also last forever. In general, too many wizards in the game with high economy but low military might break the system.

I think this should be unified to check the same statistic both for determining who are the N weakest wizards, and which of them are below 75% average. (and yes, if the weakest ones are above 75%, no elimination takes place but that's intended. If everyone is roughly equal, the game should wait until someone falls behind one way or another before activating this effect)

I'm not sure if the lowest military or the lowest overall chart strength should count however. I guess I'll try the overall lowest.

[Seravy]

One more possible thing can cause the elimination mechanic to "fail", and that is the human player being at war with the target(s). If the player is at war with an AI, that AI, while they can be an elimination target and will be attacked by other AIs, cannot retire.
So if the player never bothers to accept a peace treaty, the wizard won't retire, ever, and stays in play until the other AIs manage to banish it.

[Anskiy]

Why is that the case? Seems rather strange to me.

[Seravy]

Why is that the case? Seems rather strange to me.

I think that makes perfect sense.
It'd ruin the game for the player if the enemy they're preparing to attack randomly retires a turn before that attack.
Aside from the loss of fun, the player also loses all the artifacts from heroes and the spell they would find, too.
Of course this can still happen if they accept peace, but at least they have control over it.

Something else I found, Guardian wizards cannot retire. Not sure why that was added. Do we need it? These wizards are a fairly significant "trap" for unskilled players, as their fortress is extremely hard to take out, so getting rid of them would make the game kind of easier, on the other hand these are the wizards most likely to block an invasion from other AIs and never lose their fortress or possibly even other cities.

I've ran a few automatic test games and for most of the time the elimination works fine with the changes, however when the map had island sized continent, it often failed, as the players to be eliminated had too many cities and couldn't retire. I assume this generally is the result of any invading armies being killed on the sea... but the generic difficulty of the AI's continent management might also be part of it. If the AI already sent an army to a continent and that army was large yet too week to attack anything, the AI leaves the continent and goes elsewhere for a while. Problem is, when each "continent" only has one possible city, if the first big army to arrive isn't good enough, the AI just won't spend more and will ignore the place. Meanwhile on larger continents, that army would most likely find another city or at least a few stray units to attack and would leave the continent a valid destination for more troops.

Honestly, I would prefer the wizard actually losing instead of retiring, but the only way that can happen is if we give some artificial penalty either in combat, or economically, to the wizard, which isn't too fair and if done poorly, leads to other AIs conquering the fortress when the player sees it happen and knows it's obviously impossible for that battle outcome to happen without some sort of abnormal modifier.
Maybe if they were forced to disband units and/or run out of gold that could help, but sometimes the AI just has no units that can be disbanded in the fortress, just all heroes, summons, or even possibly 0 maintenance cost units, sanctified or otherwise. I guess if there was a better solution than retiring, we'd have that in the game instead of this already.

Edit : nevermind this last part, if it becomes easier to take those wizards out by combat, the human player can take advantage of that by sneak attacking the wizard in question only after the effect activates but before the other AI(s) manage to conquer the fortress army and just get a bunch of undeserved artifacts and a spell.
Retiring is not an elegant solution but we don't really have anything better. I'd rather the human player stole the 1-3 cities that turned neutral than 10+ high value artifacts and a free high rarity spell plus the city.

[Desertfox]

When an opponent retires, all summons and heroes vanish also, sometimes leaving the city empty and free to take, if you are quick enough. It is a pro strategy to have a cheap unit next to the cities of the smallest weakest AI to immediately take advantage of a possible retirement.
Sometimes the enemy cities may be few but well-defended.