The Next Patch

[Quickening]

Ive been away from the forum for a while and having a quick look I can't see anything about this so I'll just ask. Presumably CA are aware of this shield bug but does anyone know if it is scheduled to be fixed in this February/March patch? I love this game and am eager to get back to it but not before the shield and two handed bugs have been fixed.

Cheers! ~:cheers:

[IsItStillThere]

I think both bugs caught them unawares but they are probably working on fixing them as we speak.

I hope they find time to beef up the AI a bit, particularly in battles and seiges.

In the meantime, there are several mods that do a credible job at fixing both, give them a try.

[Laconic]

Smarter campaign AI will lead to more challenging battles, I think. I'm tired of fighting stacks composed of Town Militia and ballistae.

[Lusted]

Smarter campaign AI will lead to more challenging battles, I think. I'm tired of fighting stacks composed of Town Militia and ballistae.

That's very easy to fix with a few unit cost changes. It's not really a problem with the ai, just the cost effectiveness of each unit which the ai uses to judge its recruitment on.

[Foz]

That's very easy to fix with a few unit cost changes. It's not really a problem with the ai, just the cost effectiveness of each unit which the ai uses to judge its recruitment on.
Can you elaborate on that? Like if you know the particular formula or at least what factors count in the calculation? I feel certain that I could easily prioritize unit recruitment based on cost effectiveness, provided that I know which unit stats and attributes are actually counted in the cost-effectiveness calculation. I dunno if that's a closely guarded secret or not... but I can say that for the sake of mods we could sure use it, as it is so key to getting army recruitment right.

[gardibolt]

I'm curious about the references I've seen (I think from Lusted) to CA having identified a major bug that the community's not aware of yet....assuming it's fixed in the next patch, will CA tell us what that bug is? I'd think it'd show up in the ReadMe, but maybe they're too embarrassed....

[Lusted]

I'm curious about the references I've seen (I think from Lusted) to CA having identified a major bug that the community's not aware of yet....

I've made no such comments. The bugs i know they have fixed but cannot say are ones the community is very aware of.


Can you elaborate on that? Like if you know the particular formula or at least what factors count in the calculation? I feel certain that I could easily prioritize unit recruitment based on cost effectiveness, provided that I know which unit stats and attributes are actually counted in the cost-effectiveness calculation. I dunno if that's a closely guarded secret or not... but I can say that for the sake of mods we could sure use it, as it is so key to getting army recruitment right.

It appears to be based on the combined attack/defense values divided by unit cost. So for instance in the changes i've made(hardly major, just some refinements), militia units are now 100 florins more expensive, so the ai recruits less of them as they are less cost effective now.

[Rollon]

It appears to be based on the combined attack/defense values divided by unit cost. So for instance in the changes i've made(hardly major, just some refinements), militia units are now 100 florins more expensive, so the ai recruits less of them as they are less cost effective now.Still I face some half Ballistae/Catapults/Trebuchet, half Militia armies in LTC. Looks kinda strange to me.

[Lusted]

Still I face some half Ballistae/Catapults/Trebuchet, half Militia armies in LTC. Looks kinda strange to me.

If a faction has mianly cities, then yes you will fce some mostly milita/siege artillery units. Though god knows how the ai still manages to sometimes field 5-6 artillery in a stack in my mod, i've increased the cost by 2000.

[Satyr]

Obviously there must be more to unit selection than just a cost/specs assessment. Perhaps the AI doesn't correctly consider the number of units in a company when it does this assessment. Even Wes in Medmod 4 couldn't entirely control what the AI built and ended up eliminating some units.

[Foz]

Couple of things. First, It's a shame that they didn't choose to put the AI recruitment code into a txt file instead of in the hard code. It'd be amazing for the modding community to actually be able to tweak with army composition via a file, and all this weird tweaking to unit stats that we're doing would largely go away, or at least we'd completely understand how to make the AI recruit particular units and not so many of others.

Secondly, if we want less artillery to show up, why don't we just make it less readily available? I know I always find myself sitting on tons of artillery of all sorts in the recruitment pools throughout my empire (sheesh the AI loves to build those artillery-recruiting buildings) - so the reasonable thing to do is probably cut it down until a human player can just so recruit as much as he needs for his armies, and that should impose the same limit on the computer's availability of artillery. It's likely as it stands that the AI recruits it as often as it possibly can, so we should just make it possible less often :smile:

I'm going to speculate initially that cutting the replenish rates in half for artillery should be plenty enough to limit how much it shows up (and then I'll leave the price alone - the AI is welcome to recruit all it has available). As they appear to be 0.4 across the board, that would be 0.2, for a 1 unit every 5 turns replenishment rate, which still seems reasonable once you remember that many places will be able to recruit 2 or 3 different kinds of artillery units, each once every 5 turns.

I'm not sure across the board cuts are actually needed though, perhaps a more surgical strike is in order. Can anyone give a good representation of which units they typically see as the offending artillery units that are overproduced (if it's any specific one(s) as opposed to everything)? I haven't encountered this one enough times to generalize about what exactly is wrong...

[General Zhukov]

Couple of things. First, It's a shame that they didn't choose to put the AI recruitment code into a txt file instead of in the hard code. It'd be amazing for the modding community to actually be able to tweak with army composition via a file, and all this weird tweaking to unit stats that we're doing would largely go away, or at least we'd completely understand how to make the AI recruit particular units and not so many of others.

Exactly right Foz. If possible, I would urge limiting AI armies to two pieces of field artillery per army, max. Mostly I see ballista abuse on the computer's part. Meaning they bring along too many and can't figure out how to use any of them. Sure, a player, if skillful and careful, get some use out of the things, but the computer has no clue and treats them like weak catapults. So, I would like to see a mod that does away with ballistas and caps the others at two.

[Ars Moriendi]

Couple of things. First, It's a shame that they didn't choose to put the AI recruitment code into a txt file instead of in the hard code. It'd be amazing for the modding community to actually be able to tweak with army composition via a file, and all this weird tweaking to unit stats that we're doing would largely go away, or at least we'd completely understand how to make the AI recruit particular units and not so many of others.


Sadly, CA is not in the market to sell game engines, but fully developed and featured games.
As such, there must be a limit to what we can do, compared to the software developer : no source code for us, only parameters and settings to tweak and artwork to change. Anything more than that, and we will have to look much deeper into our pockets for licensing the engine.

As for the artillery spam : does the game looks at the melee stats too, when evaluating cost effectiveness ? if so, lowering att/def for arty (useless anyway) might help.