Sawmills

[Seravy]

Escalating the above idea.

New outposts start with Smithy already built. Smithy produces the two basic hammers. Smithy has a higher (125-150) cost. Fighter's Guilds cost 80 more. Smithy has a higher than normal chance to get destroyed by building destruction effects (battle, conquest, etc).

Pros
-Conquest becomes less desirable. If the smithy breaks, you end up with a hamlet that's worse than a newly founded one (missing basic production and ability to make swordsmen. So you need to do housing for 1 more pop to make up for lack of 2 hammers, and need to invest ~200-300 if you also want to be able to produce units.). (Unfortunately, if the AI bought marketplace or other basic buildings, you'll still be looting more buildings than in the previous system)
-AI can get those basic swordsmen faster
-Swordsmen tier unit might get to see a bit more use : you can build them asap if you wish, no need for a cost 80 building, and cost to reach halberdiers is more.

Cons
-Spearmen even less desirable if swordsmen available immediately (but they are meant for spellcasting enablers and unrest reducers anyway, not real troops)
-AI will be several turns more behind if using the "economy first" strategy because the time to build up is less, so less chance to "catch up" and break even with the housing delay. With FG costing +80, total cost remains unchanged for the AI.

I can't really find any reason why this wouldn't be a major improvement, anyone else can?

[Nelphine]

I'm not sure you need the increased destruction on Smithy, unless you specifically intend to make it very hard to get any military buildings from a conquered city (since all the other military buildings including stables/fantastic stables are based on it.) But if you do that, I feel like you should also increase destruction rate of wizard's guild and shrines and parthenons, otherwise, you just end up making magicians/shamans/priests way more common.

Personally, I didn't understand the reason to increase the cost of the smithy at first. But, then with the assumption you want to make it harder to get conquered cities going, I think it makes sense. But in that case, the increased destruction rate and the increased cost of smithy are doing exactly the same thing - so I wouldn't do both.

So I'd remove the increased destruction rate, and leave the increased cost.

However, by adding smithy as your free building that gives a city free production, you haven't actually increased the overall production of a city, because of maintenance of the smithy. At the beginning, a city prefers hammers, but later on cities prefer gold. So this change in conjunction with decreased hammers from sawmill will reduce the productivity of the empire. Not saying don't do it, but I think it will hurt any race that doesn't have access to miner's guilds or mechanician's guilds. (I think High Men are going to end up overpowered due to the decreased building cost by the way. It's simply too huge a bonus. Although they won't be able to use the new housing as much I suppose.)

[Seravy]

Good point about maintenance, we should remove that.
Albeit it doesn't directly interfere with any of our goals. We needed hammers specifically, losing 1 gold/turn is not relevant (and there is no ROI associated with basic production/free smithy). Yes, that means we are at a net -1 gold/turn unless the city would have built military, but since we have a more robust start on economy (higher ROI housing than sawmill/marketplace etc for the first few turns), that maintenance is probably paid for by that, albeit it does make the mid/late game worse.  So yeah, to be on he safe side we should have 0 maintenance on Smithy.

Increased destruction of smithy would stop being relevant once Fighter's Guild is built - you can't destroy something that's replaced. So that would be specifically an early game feature. As a side effect, it would make combat damage done to cities owned by human players (who typically do not build fighter's guilds everywhere) more severe.

I still have no good idea what Builder's Halls should be doing instead of boosting housing.
...does it even need to do anything? It unlocks Engineers (where available) and advanced buildings (where not).

Maybe it could reduce the population destruction rate caused by combat by, say, -5%? (Strictly only the population, not the buildings.)

[Nelphine]

I'd be fine with not having it do anything; it's really just an extra 20 production before getting things like Parthenon, contributing to why they are cheaper (the first advanced building is more expensive). Adding anything just makes those races that get it have that much more advantage.

[Seravy]

Started working on it.

For housing, "max pop" will be the result from the "terrain max pop" procedure. So Granary, Famine, Wild Game not included. But Gaia's Blessing included. We don't want to encourage building granaries prior to housing (would have a pretty major effect on bad locations, turning them average).

I think the "everyone to workers" effect will need to go into the "minimal farmer required" procedure to ensure it's not causing food shortage and it gets re-enforced every time the player tries to change farmer amounts. Sawmill, Smithy changes done. Currently working on housing. AI changes, easy lairs, will be done after housing.

Would be nice if I could put a description of the new formula into the game, but while Housing does have a help entry, it is not being displayed anywhere (you cannot click on a "housing" building since it's not one). The usual short building description does not have enough space. So I guess I can at best change it to mention it requires workers, no room for formulas.

[Seravy]

Ok, this forced worker thing is a bit problematic.
The AI sets tax/food first, then production. Then it won't have enough farmers if it sets a city for housing. Which makes them disband troops next turn.

So we'd need to swap this order - set production first, then tax/food. I don't see any particular problem with that, none of the building decisions are based on taxrate nor food production. The AI would also need to learn to never set farmers in cities that are currently housing.
Where I do see a problem is, if the AI has enough troops that they need the farmers everywhere, and switch to housing. But that shouldn't realistically come up ever (only the smallest cities will build housing, if the AI has no larger cities, why would they have a lot of troops?) - to begin with the population set to workers in the housing city shouldn't have even existed before. 

Or, we could avoid this whole mess and say this restriction is human only.
-The human can't see when the AI is doing housing, so even if they see a farmer they wouldn't know it's illegal.
-The AI never allocates more than necessary farmers, so extra food wouldn't be used to produce gold to get around loss of hammers.
-But the AI would get extra hammers by having farmers there which means they aren't having them elsewhere so that elsewhere is getting the extra hammers. So it's not really fair.

So I guess we need to do it the hard way...

Wait, if the human sets the city to housing in a way that does not immediately check minimal farmers (like the cities screen wouldn't I guess) then the "all workers" isn't enforced, they get to end the turn, then the farmers change and they lose units.

I see no good solution for this, any ideas?

...making housing have reduced effectiveness if there are more than the minimal farmers is something I can do, but then the AI needs to learn to not set farmers there...wait I can do that.

[Nelphine]

Ugh that sucks. I really think it would be better to maximize farmers - otherwise it's yet another thing you really need to display in the description of housing.

Maybe see how much you can fit in the description (you don't need the whole formula, but you do need to explain what works and what doesn't), and see if you can match that to something you can code?

Otherwise, housing will become something you need to look up online to know when to use (and much worse, how to maximize it), which I think is a bad idea.

[Seravy]

Well, I did manage to squeeze the most important basic information in : what increases and reduces efficiency.
For actual numbers, they'll have to experiment or read the changelog.
Now doing the AI changes, then I'll need to test the AI, and after that only the easy lairs left (which is probably the most difficult) but I suspect that'll be left for tomorrow.

[Nelphine]

Yup that's all that really needs to go in. That should work then.

[Seravy]

Also since the AI now has a way to spend gold (It learns to spend it prior to housing on marketplaces and magic markets) that's highly relevant, we probably should give them the 500 gold we are giving the player in lairs.

Except, they aren't really needing it.

Ultimately, only the first ~400 gold matters (which the AI can use on their marketplace and magic market purchase in their capital on turn 1), the rest will have minor early game effect and will eventually cause the AI to build random things a turn or two earlier than otherwise until the gold is used up.
...which kinda means the rest of the gold isn't necessary, but it being there also means the AI will have a better chance of hiring nice mercenaries, or heroes, making the game more interesting.

So yeah, I guess it doesn't really matter how much gold the AI gets, as long as it's at least 400. (by the time the outposts turn into a city, the capital's marketplace plus taxes will pay for the marketplace in the other two cities anyway)
Looking at the current difficulty table, they get above 400 on Expert or higher, the seriously difficult levels. On lower levels, they get less, but there the human will also be less efficient at exploiting their free lairs.

So overall I still think it's good as is and needs no change.

I do wonder however if the AI should be trigger happy on buying sawmills (which risks not having money for a much more important pop 1 marketplace/magic market in the cities founded later) or not (which means their capital will waste ~2-5 turns building it but they get to save most of the gold). Let me see, if they spend about 10 turns on their outposts turning a city (and settler reaching the tile to build), and have marketplace bought, they earn 8+some from taxes, so about 10-12 gold a turn, plus AI bonus. So on expert they'd roughly have enough for one other marketplace, but they'd need enough for four (two cities*two buildings). So better to save the money? Getting one more marketplace 10 turns earlier is better than the capital being 2-5 turns faster on their settlers, etc? (That marketplace also pays for another marketplace after 20 turns...)
Yes, I think that's correct.

I actually wonder if the change benefits the AI or the human more. The AI can definitely snowball better from marketplaces funding each other, and it's likely also smarter at knowing when to switch housing. (Yes, I gave it the entire table of data so it stops at the exact maxpop+race where it drops below 15 ROI)
On the other hand the human simply having the option to have gold to invest compared to the old system where they had none but required it for the sawmill, helps the human tremendously (or more like, removes a huge disadvantage on their side). (and in comparison makes luck with treasure and sprites less powerful than before - it's kinda like as if you always played a game with an average or better start. Which was the goal in the first place so yeah.)