The current update broke this. It will remain broken until I have time to go over the changes to recruitment priority pointers.

Update the mod to the current version

sadly starting the game with this mod enabled is not possible (only me?)

Nothing eliminates that. The AI will always recruit lords too quickly and thus have at least one army sitting out in the open recruiting.

This mod eliminates AI STANDING near its cities ?

Going to have to completely redo TMS from scratch and partially remake the AI personalities, so it's probably going to take a while.

Thank you for the tentative timeline, really looking forward to the update!

Due to changes in the underlying AI algorithm, it looks like I may need to start AI personality work over from scratch.

ETA: 1-2 weeks.

The AI is better at determining garrison strength, particularly on higher difficulties. That won't stop the AI entirely from attacking very strong garrisons (such as a T5 landmark settlement in my own garrison overhaul), but it will usually dissuade it unless it can muster enough troops such that it thinks the fight is winnable (whether or not that is true is up to your modlist and your skill). Now, if - for some reason - we are talking about vanilla garrisons, I must comment that that is a strange question on account of vanilla garrisons being laughably weak.

Finally,since the interaction between the task generator and threat detection is probabilistic in nature, your mileage may vary.

Does the AI suicide into garrisons with this mod? I've tried every other AI mod (ultra aggressive/smart etc.) but the AI still decides to suicide into garrisons for no reason :(

The former is likely an example of a victory-flagged settlement being offered for trade (the inverse is that factions with flagged settlements won't trade those settlements). The latter is doubtless a product of war balance and/or the "last stand" diplomatic weight modifier. Probably both, as Thorgrim is routinely at war with nearly everyone in the World's Edge Mountains.

At turn 8 or so, i offered to trade a tier 1 minor settlement to newly met major faction. They were a bit stronger than me, and i was their (-3) strategic threat. That offer made them so happy, they became my vassal, signed trade, gave me all their bank (12k) and there was still ample room for further demands. That settlement trade offer was green at +192.

In the same spirit, playing as thorgrim, i could confederate any dwarf faction at any time, even if all other deals would be deeply in the red.

Maybe the AI is a bit too much troubled or fearful when it has few settlments?

Hope it can help for future development, should you decide to do so :)

Thanks for always being helpful. I appreciate it!

Yeah, steps were taken (by way of brute force) to attempt to force more agents to embed into armies, wherein they can actually be useful. The results were not as effective as I'd have liked.

That said, that should be the correct table. Note that the values are percentage per-turn chances of tasks being generated, that each faction uses multiple profiles, and that said percentages are cumulative. As such, altering the default profile ought to have the most noticeable effect. There's also a task limit, but that shouldn't matter.

I think I found it under db/cai_task_management_system_task_generator_groups_generators_junctions_tables. Woot!

Can you point me to the part in the DB tables where I could make the AI be a bit more aggressive with their agents again? I've tried looking on my own and it is all greek so far. Thanks.

I hope you do come back and find the cause :)

Vanilla AI building priority is abysmal and playing with your mod was a big change for the better.

Again, I'll look into it once I play WH3 again. This is starting to sound more and more like vanilla AI behavior if you memory edit it into severe bankruptcy. But that shouldn't affect TK's because, as has been said, they have no recruitment related costs. If I had to guess, this'll turn out to be something budget allocation related.

TKs in my game went completely passive. It looked more like a freeze of any activity than just recruiting. Just lack of recruitment was the most obvious since TK never have reason to not recruit (no upkeep). Before that I had to fight a stack from them almost every other turn.

I've got the stated issues in the bug tracker, so I ought not forget by the time I come back around to Warhammer. Hopefully, by then, CA will have fixed whatever the fuck is wrong with the diplomacy hint tables that causes crashes when I alter them.

@Commisar Jon Fuklaw thanks for offering to look into Bretonnia's farms. I hope you have a great new year!

There should be a marker for mills, but it's entirely possible there's not. Even then, that particular building synergy probably requires a contextual hint anyways to prevent overrides from the construction personalities. Will look at it Soon™.

As for TKs not recruiting, that's odd given how high recruitment priority is for them. Again, it's an issue for Soon™ (or Soon™ after Soon™ since I'd need to replicate it), since I'm not playing Warhammer at this time.

@Lycia I actually play a Breton campaign now and I can confirm they don't build farms up - and the reason is simple - to upgrade farms you need the mill building and AI never builds it.

One more thing that happened in my campaign - a tomb king faction that for a few turns sent stack after stack on me at some point broke. They just stand in place, don't do anything. They have only two armies (turn 97 they should have at least a few more), they are not even full and they do not recruit.

Again, that one's not something I've seen before myself, but it could be a result of other mechanics (IIRC, the brettonian ecnomic setup is weird). I'll double check later to make sure I didn't put a limiter on the chain, but it shouldn't have one, as it's a non-military building chain. Worst case scenario, an additional contextual building pointer can be added to the higher level buildings in the chain.

I'm taking a stab at setting the "Building or building range start inclusive" tag to wh_main_brt_farm_1 the end inclusive to wh_main_brt_farm_3 to see if that does anything.

This may have nothing to do with this mod as I use many, but I am noticing that AI Bretonnian factions in my campaigns are not upgrading their farms past tier 1. Is it possible the build priority on the t2 farm is wonky? From my reading of the pack file it seems to be the highest priority of all AI income buildings, but I figured I'd take a shot in the dark.

I'd say I aim to please, but that's not really true. Either way, if you have further questions, I can likely provide useful answers.

You've gone above and beyond with my question. I gave you the steam point clever award on this mod! You've earned it! =)

If it's a desperation move by the AI that's the issue, then dropping the last stand factor for OFFER_JOIN_CONFEDERATION in cai_personality_deal_evaluation_deal_component_values by a factor of - say - 10 ought to prevent the issue, but it will make saving LL's significantly harder.

yeah, am am using a mod to enable vassalization, then I have fudging vassals which has a ton of neat fixes but it also allows you to confederate a faction immediately if you vassalize them and are the same race, which leads to the cooldown avoidance. I'll figure out a way around it or just not abuse it in future campaigns cuz I like all of these mods lol. I have like 144 active and I edit local copies of tons of mods from the workshop trying to make the game play like chef's kiss (in my opinion, others would see this as an abomination lol.)

All of that said, the High Elves are typically all rather tight knit now that they have multiple "dangerous" (as in likely to trigger the "Last Stand" diplomatic bonus by merely existing) threats, so if you are allowing confederation of vassals and have the cooldown and/or penalty disabled via other mods, then I'd call that a likely source of the "issue".

Thanks. Given that my issue combines an interaction between three mods moreso than any one of them, I'll look into how you and the other modders set the numbers up and tweak accordingly. Thank you for taking the time to reply with useful information! Have a great one!

As for how any of that interacts with other diplomacy-related mods, since all of those touch AI (except those that simply enable occupation/dilemma options) technically conflict with this. One or the other will come out on top in terms of what manifests in-game. Which that is should still be based on table naming conventions and which table has the higher priority, but I've not bothered re-confirming that for WH3. If so, then odds are my tables have the higher priority, as only one key has higher priority than @.

The weighting curves for many diplomatic agreements has been altered, including confederations. Typically, any faction that makes it to Trusted Friends has a very good chance of accepting confederation and, if not, will be within 50/50 odds by threatening them.

Unfortunately, the second order attitude adjustments that were in place to make getting to Trusted Friends significantly harder in most cases were removed, as apparently something broke with that table in the last patch and I finally isolated it to any changes being made to that table a couple weeks back.

If so, when combined with the Fudging Vassals mod which lets you vassalize as races that normally can't and offers to let you confederate new factions, it's possible for example to unite ulthuan 80% by like turn 20 (no confederation 5 turn cooldown). If you think the mod is tweaking it to be easier, just lemme know and I will edit the local copy instead of asking you to fix an issue that arises partially because of another mods.

Is there a chance this mod is making it easier for me to vassalize/confederate factions? One of the mods I use is making it too easy and I am trying to pin down which one.

Dwarfs as a race are flagged for more defensive play, both through Task Management and via their construction/recruitment/diplomacy profiles. That by no means should make them fully passive, so it's odd you are reporting that. Karaz-A-Karak is frequently the largest non-skaven faction in terms of territory when I see them.

It honestly sounds like they are losing the early fights and not using confederated forces to open new fronts, which may be a function of autoresolve balance (whether that be the result of vanilla, piecemeal modding, or a major overhaul).

I can tone down defensive tasking for Dwarfs Soon™, but given all of the interactions involved, there definitely is the concern that the next thing I hear is "OMG THE ENTIRE WORLD IS COVERED IN BEARDS! DORFS OP NERF NOW! etc etc etc".

I've noticed doing some playtesting on observer games that the Dwarfs are extremely passive. Is that intentional? Karaz-a-Karak and Karak Kadrin in particular seem to want to turtle in their one home settlement and never leave.

Aye. The Task Management System weights heavily favors local expansion for all AI factions. If you are adjacent to one or more AI factions and they hate you...well, expect visitors. Conversely, well fortified provinces won't often be attacked if sufficient forces can't be allocated to overcome expected autoresolve outcomes. Note that this only matters if you are using something that undoes the severe nerfs to settlement defenses, as the AI can run roughshod over vanilla garrisons.

Does this effect Anti player bias?

UR-025 is a galactic treasure.

I like the Men of Iron being the thumbnail.

Haven't seen anything about that. Will look again.

colonisation now costs no manpower and instead gives a turn long heavy vigour debuff to the colonising army, just a heads up in case u missed it

Louen shouldn't be razing anything. Oh well, time to change a couple of TMS pointers.

Starting as Be'lakor, I see AI Louen raze the minor settlement in his province every time, only to colonise it a bit later. Is there any hope to make them less dumb in this regard?

So long as that mod doesn't rely on non-military buildings for unit cap effects, it's probably fine. The AI will likely suffer, however, as there're pointers to prevent building duplicate military buildings in each province, and unit caps require building duplicate buildings for the sake of raising caps.

If the author is smart, he'd have accounted for that and made his own changes to the construction management system. At that point, it'd just comes down to table names to determine whether the AI is fine in that context or not.

would this be compatible with the Unit Caps For All mod? it turns all factions into have the same unit cap requirements as Tomb Kings do and requires buildings to increase those caps. Not sure if that has an effect on ai building priorities with this mod

I started on some additional diplomacy work but never got finished; got sidetracked. Should have time this weekend to actually do Something™.

Gonna turn this off for a bit until you patch this up for 2.2 and we see how the new changes pan out in general, but I gotta say, I was impressed.