Todo list for 5.44

[Nelphine]

No, right now, rush can kill the last ai before very rares (before rares if you really push it). The game design is roughly that each of the 4 AI should correspond to a tier of spells. There's a reason the max bonus for time taken to win is awarded at 1420. The game is literally not meant to be won before that.

It doesn't matter if the AI are 10% higher in boosts by 1410, the rush will kill them anyway. But if you manage to actually follow the game design (and I suggest, you've never played a strategy that came close to meeting the game design), then the last ai, even second last one, would be MUCH too strong with a 10-20% further increase over what they have now.

So, no, progressive wouldn't do anything, unless you make it so high, that the ai can't be beat by rush. But on a linear scale, that means you'd need, what, 100% more by 1410? So proper long term strategies (which means winning no earlier than 1420), would be 200% stronger . That's unbeatable. 

So no, progressive does nothing to help against rush, unless it's so strong that winning simply becomes unbeatable.

There are other reasons for progressive bonuses. Solving rush is NOT one of those reasons.

[Seravy]

Yeah I don't really see why the late game would need to be harder. I've only ever seen or played one of four kinds of games :

Rush which won too early and there was no late game whatsoever, game ended in like 1408-1412, but ultimately was won in 1403-1407 already.
Rush which failed to win early (preferred outcome by design), and acted like a late game strategy afterwards to take advantage of conquered territory obtained by rushing. The advantage of having that territory balances against your other disadvantages, if you even had any (werewolf rush uses 10 books, it doesn't suffer from lack of spells!) so ultimately these turn into "normal games" past 1408-1412, so see below :
Normal games where conquest happens at the designed pace, in these the AI always managed to be difficult (in fact usually more difficult than expected) and quite often actually winning in the late game (looking at you, Myrran wizard with Death books)
"Peaceful" games where pure economy is the deciding factor, these were actually extremely rare (hoping to see them a bit more often after some of the recent changes) and generally win if they manage to obtain lasting peace with all AI, lose otherwise. Regardless of how much resources the AI is getting, as they either reach their win condition the AI can't hope to stop (like Time Stop), or not.

None of these would benefit from, or need a progressive AI bonus.
(also I'm yet to actually see someone other than myself play a high difficulty long term, non-rush game and win.)

Either way, think about it.
If we want to encourage players to play longer games (to fully enjoy the higher tier spells and units), we should make it harder to defeat enemies early and easier late, not the exact opposite.

I don't think I can say anything else meaningful after all the above, so I'll try to avoid commenting on this from now on.

[zitro1987]

I tend to struggle during late years (expert): the movement of multiple stacks to your units or cities or nodes. Enemy AI also appears to improve their power and skill levels at least as fast as I do. Their spell usage involving more advanced spells can also be terrifying at times. It requires more brainpower than the early game for me - the early game I'm learning how to properly have sufficient troops to avoid war.

In short, late game doesn't need to be harder. It already requires a lot of brainpower to play it, even when you are ahead.

[Bahgtru]

I give up. I have never proposed that this would solve rush btw.

Given the differences, please at least don't make the game less challenging by reducing the difficulty (and AI cheating bonus) at lunatic, when you do the rebalancing that you've spoken about. Keep it like this for whatever reason - finding overpowered strategies would be a good one.

[Domon]

just as a brainstorm: can the AI bonuses be made adaptive, i.e. make a portion of the bonus inversely proportional to how much they are behind or ahead of the player?

could a single difficulty setting be made to work like this?

[Seravy]

eww that defeats the purpose of having difficulty levels.

But either way I've said many times, variable AI bonuses are not possible. It's an array of constants in the data segment that's being read in many different places in the code. At the very least, not possible for the types of resources where the data is used in many places, like AI overland casting. Others like gold might be doable but this should be an all or nothing type of feature.

[Domon]

"eww that defeats the purpose of having difficulty levels."

actually not

2 scenarios:

1) this becomes a separate difficulty level; other difficulty levels are unchanged, nothing is defeated
2) this become a general feature: the starting bonus and the scaling can still ben made more or less favorable to the player on different difficulty levels

"But either way I've said many times, variable AI bonuses are not possible."

fair enough, wasn't aware of that

[Seravy]

Finnaly finished my ongoing test game and started doing some real work.

So far done "1.", halved Stoning chance and made item types weighted in a ratio of 2/2/2/2/3/3/8/2/2/2 (8 is accessories, 3 are staves and wands, 2 are everything else).

[Seravy]

Finished implementing these so far :

-Chance of Stoning to appear on random generated items is reduced. Due to its low cost, it was appearing excessively often.
-Random generated item types are now weighted : 2 Swords, 2 Axes, 2 Maces, 2 Bows, 3 Staves, 3 Wands, 8 Accessories, 2 Shields, 2 Chains and 2 Plates on average for each 28 items. Original was 1 of each type, which considering 6 types of weapons, 1 accessory and 3 armor was quite inbalanced.
-When a spell would be found in treasure but no spell of the appropriate tier is available, 200 gold/tier is added instead.
-If the human player is playing 4 or more Death books, AI will prioritize researching Exorcise, Holy Word, Cockatrices, Shadow Demons and Petrify more.
-Fixed bug : When a settler is produced from the last unit of population and causes the city to disappear, the production cost is subtracted from a random other city.
-Werewolves now have 4 resistance.

So let's see the list now.

1. Done
2. Will do if space allows
3. Still considering
4. Will fix
5. Probably should do, not sure how though, will need to check if it's even possible
6. Done
7. Done
8. Not doing this time, might consider if werewolves still too powerful after the update. Hoping for 6 and 7 to solve the problem.
9. Definitely doing if able, haven't investigated yet
10. Probably no space for this and it doesn't really improve the game anyway. Seeing a tiny chance we might disable ores on werewolves later if nothing else can fix their balance as well so postponing this one.
11. Need to investigate if possible or not.
12. Will keep at -2 for now, Demon Lords felt good enough in the game I played (not the last one but the one before)

And some more things I considered meanwhile

13. Since food buildings were redesigned, how about making Nomads build Granary but not Farmer's Market and Forester's Guild? That way they have even less food than now, but still get the +5 population. While having a race that has 5 less max pop is nice for variety (and makes those 21-25 pop places relevant), it really contradicts what their racial bonus is for (trade bonus scales with population. You need at least 17 population to use all 50%.). However this also makes them no longer vulnerable to city curses (in particular Drought). Also, less food doesn't really work for Nomads either - they are the only late race that actually gets a powerful early unit.
I feel we definitely need to split the "5 less max pop", the "trade bonus" and "Produces less food" traits between races, they are not really compatible. Also the race has too many special traits as is, they have "gold bonus", "less max pop", "less food", "more resistance", and even "bow on normal units".
I think what would suit the race best is to keep the resistance, and the less max pop as they are a package (strong against Death in battle, weak to max pop reduction curses). Trade gold bonus is quite needed for the race despite the max pop problem - they are a late race with no engineers. It also suits the flavor of the race. So probably keep that, too. It is the only decidedly bow based race (Rangers, Horsebowmen as special units) so that has to stay as well. Less food unfortunately can't be avoided - no granary means no farmer's market. And with this train of thought we are back to "keep as is".

14. Do Demon Lords need to be immune to petrification? It's not that it actually works on them much with their 13 resist, but in certain rare scenarios (cockatrices, or black prayer) it might. I'm leaning towards letting them keep it though.

15. I don't remember if we had a discussion on "Death can't hurt other Death units in battle" before? While it's good for flavor, it's not that good for gameplay, staring at my spellbook "oh, none of these do anything for this battle". On the upside it encourages playing multiple realms, which is nice as pure Death is kinda above average right now. If we want change, enabling 1-2 spells to affect Death units would be enough.

[Nelphine]

I'd be inclined to make ghouls not have standard death immunities, though they could still have some, and the same for werewolves. That way in common and uncommon, death spells could still be relevant. Then make .. annihilate? The one that kills one unit - ignore death immunities, so very rare has something. That leaves just rare, I'm not sure what to do there.