Does the AI ever exempt provinces from tax? (When it's low on food)

[Fridge]

Just wondering about the food issues the big AIs seem to suffer, leading to the collapse of big empires and subsequent lack of challenge for the player...

We've all had food problems (haven't we?) - when conquering a couple of territories coupled with a sudden building splurge suddenly puts your food into the red. But that's not a problem, usually by this point you're making thousands a turn so it's an easy matter to exempt the biggest food eating provinces from tax (and thereby taking them out of the whole food supply equation). Just wondered if the AI was programmed to do this, and if there was any way of checking if they ever did do this? Might explain a lot?

[Bramborough]

Just wondering about the food issues the big AIs seem to suffer, leading to the collapse of big empires and subsequent lack of challenge for the player...

We've all had food problems (haven't we?) - when conquering a couple of territories coupled with a sudden building splurge suddenly puts your food into the red. But that's not a problem, usually by this point you're making thousands a turn so it's an easy matter to exempt the biggest food eating provinces from tax (and thereby taking them out of the whole food supply equation). Just wondered if the AI was programmed to do this, and if there was any way of checking if they ever did do this? Might explain a lot?

It's an interesting question, and I don't know the answer. My observation, however, is that for the vast majority of factions this wouldn't even come into play, as they rarely have more than a full province's worth of territory (and even then often not unified within an actual province's boundaries). A 7- or 8-settlement AI faction is unusually large (and rare). So if they were to exempt a province, that would be like half or more of their entire territory.

I'm not saying it's not a factor for the very occasional large AI (some of the far eastern factions do get pretty big sometimes). But the AI has problems much more fundamental, which if they were solved via patch, would benefit small and large AI faction alike.

[phred]

In one of my campaigns I took the Syria province from the the Seleucids.
I noticed before I attacked that they were having public order issues, and after I conquered the province I found out why.
They had upgraded most buildings to the third tier. The province was making a ton of money, but the squalor was out of control.
I tried to balance it but ended up tearing everything down (and reduced the province income by half or more).

I think the AI has trouble balancing income, food, squalor, and public order and probably doesn't have the ability to tear down buildings and re-balance things.

[Sociopsychoactive]

There is a way to monitor this in theory.

A spy close to a settlement will tell you all the pertinent information about it, relative to the province it is in and this includes income at a settlement level, and a province level on the province tab (though this is hugely innacurate more often then not if you don't have a clear set of information about all settlements in the province, and as the AI never gets big enough for corruption to be a major issue it is usually much higher than you will actually get.)

These numbers are pretty accurate at a settlement level and, in theory at least, should therefore tell you if the province in exempt from tax. I have never seen the AI do this at all, ever.

In responce to Bramborough's point about AI and provinces, I disagree. Carthage, for example, has loads of settlements that are the only one in that province they control at the start and ,if coded correctly, could happily exempt one settlement without losing a large amount of income and have that settlement still produce the right troops ect. This could work, but ATM the AI doesn't use it.

[Bramborough]

In responce to Bramborough's point about AI and provinces, I disagree. Carthage, for example, has loads of settlements that are the only one in that province they control at the start and ,if coded correctly, could happily exempt one settlement without losing a large amount of income and have that settlement still produce the right troops ect. This could work, but ATM the AI doesn't use it.

I would agree that that Carthage is an exception due to its decently large/rich but fragmented starting position (which may be one reason why AI Carthage usually fails pretty early in the game...too many sea-separated settlements to adequately defend with starting forces). Other island-start minor factions such as Knossos, Syracuse, or Rhodes may also have a higher tendency to acquire single settlements in a number of different provinces. In general, however, most factions are a bit more contiguous. Hence my use of the term "majority" rather than "all".

[Sociopsychoactive]

I would agree that that Carthage is an exception due to its decently large/rich but fragmented starting position (which may be one reason why AI Carthage usually fails pretty early in the game...too many sea-separated settlements to adequately defend with starting forces). Other island-start minor factions such as Knossos, Syracuse, or Rhodes may also have a higher tendency to acquire single settlements in a number of different provinces. In general, however, most factions are a bit more contiguous. Hence my use of the term "majority" rather than "all".

Fair point and well made, but again I respectfully disagree. The norm of very small factions is indeed true, and for those you are indeed correct, exempting would cripple the economy and the whole faction, but the factions that expand beyond that tend to follow a similar pattern, and tend to be a similar group of factions, one of which dominates the others. My examples are all from the south-west mediteranian, but similar paterns are seen in other area's when playing as other factions.

Syracuse has very often risen to become a medium player in my none-roman campaigns (I usually wipe them out as rome pretty early on). They have struck out pretty randomly and while they often end up with one complete province, the rest of their holdings are almost always one settlement to a province.

Sparta often grow to be a medium power (I would normally call them major, but almost no empire ever gets large enough to be a major power under the current AI) and again, while it may well hold two complete provinces, it's gains further afield are almost always one settlement of a province.

The other allies of carthage (Carthago Nova and Libya), who more often than not outlive the master faction, also not only start with but tend towards conquering a province or two at most and nibble at the edges of other provinces, conquering a settlement or two.

While all of these examples are sensative to circumstance, they happen often enough to be the norm rather than the exception, and a none-roman game will tend towards one of these three becoming as large as any other AI faction in the game sooner or later. In these instances the faction would almost always benefit greatly from exempting a province from tax sometimes to relieve food shortages.

The flaw here however isn't the AI not doing just that, the flaw is the AI STILL not building in a reasonable way even with Patch 4 targeting this specifically. It will consistantly build minor towns up to level 3 and 4 (presumably for the income boost) and end up with major food problems, or at very least not having a reasonably food surplus. It then builds high level military buildings for the troops. Normally this mix would lead to public order problems, and in Bramborough's norm this isn't an issue as the AI park it's main stack in it's only province and the public order issues practically go away. Bad AI prioritisation of buildings is the problem still, and this would be a sticking plaster aid at best.

[Bramborough]

Fair point and well made, but again I respectfully disagree. etc...

Lol, ok, we can disagree on this rather tangential point. I sense the difference is that I'm considering the entire range of factions, most of which remain quite small (say, 3-4 regions or so), whereas you are thinking in terms of that subset of factions which attain sufficient size for a "tax-exempt" AI capability to matter. I am willing to concede that the proportion of these "medium"-and-larger groups may be higher than I've assessed. I'd have to go through a couple of campaigns bean-counting regions every few turns to come up with some data, but who feels like doing THAT...it's a game. In any case, I agree that ability to tax-exempt a province would indeed help sufficiently large AI factions...if the AI were capable of using this ability in a coherent manner. Currently, I don't see any evidence that the AI has such capacity.

Quibbling over secondary point aside, I believe we are in firm agreement about a much more central and fundamental problem: The AI doesn't build very smart and remains incapable of balancing food/squalor/income. This issue afflicts all AI factions, regardless of the size or configuration of their territory.

I do believe Patch 4 has made some progress, particularly in earlier build decisions. In early game, I've captured a few same-culture settlements to which I made very few changes...and even those only because I had specific ideas about that province's future role in my empire, not because the AI had built it economically unsound up through the Level II buildings it had available. It's when Level III/IV structures introduce squalor and food costs that the AI still goes awry. So my impression is that the AI, with Patch 4, is starting to make better decisions (at least sometimes) about what types of structures to build...but is still incapable of staging them up past Level II in a coordinated manner. The tech tree is the AI's worst enemy. More work to be done, CA!

[phred]

The AI doesn't build very smart and remains incapable of balancing food/squalor/income.

Don't forget the public order buildings like temples - food/squalor/income/order

I wonder if the AI would do better if it was limited to building Tier 2 upgrades. Or maybe a combination of Tier 2 upgrades and Tier 3 barracks.
A tier 3 barracks would give them good troops, and they could balance the food/squalor cost with tier 2 upgrades.
Then maybe the AI would have some serious empires.

[Bramborough]

I wonder if the AI would do better if it was limited to building Tier 2 upgrades. Or maybe a combination of Tier 2 upgrades and Tier 3 barracks.

I've often thought same thing exactly, may have even posted something like it (too lazy to go hunt). If restricted to Tier II, AI would do much better. Heck, I often hold my own provinces at Tier II until I've got everything unlocked that I need to balance; the right temples, etc. And then do the same at Tier III before proceeding to IV. I think the AI simply builds each individual slot to the highest tier possible for whatever building line occupies it, without regard for other slots/techs needed to compensate.

During first campaign, I did a lot of stupid stuff trying to get the hang of all this, and did a LOT of tearing down and starting over. I'm sure the AI doesn't learn from its own mistakes in such a way, ultimately just does what its programmed to do. However complex various if/then statements might be, it's bound to go off-track at some point and be unable to recover. A simple "don't build above Level II!" rule would leave it significantly underperforming a human player....but still far better than it manages now.

The more I think about this, the more I realize that "death-spiral" AI factions don't seem to appear until the campaign has been going for a while, and tech progress has had time to develop. I don't think I've seen very many starving-waif factions in, say, the first 50 turns or so.

As far as barracks, I've only played Rome and Pontus thus far, and am not sure what peculiarities other factions may have in their troop progression. For these two, however, Level II would be just fine for some pretty capable armies. Veteran legionaries, IMO, are the highest troop quality than Rome needs to complete a full campaign (it's certainly fun to field the higher ones though). Pontus gets a perfectly serviceable infantry core (hoplites) even lower, at Level I. I haven't upgraded to their Level IV Bronze Shields yet, but am pretty confident that the hoplites would get me all the way to the end if I so elected.

My point being, I'm pretty sure AI could build some robust opposition armies if held to Level II barracks alongside other Level II buildings. Certainly we would prefer that the AI could competently build to Level IV and not only recruit but maintain the highest troops in the game. Right now that doesn't seem possible. Better that they give us some healthy Level II opposition than some attrition-starved Level IV remnants.