1.2/1.3 Community Buglist

[Rozanov]

AI still not behaving very sensibly in 1.2 with siege artillery.

I was defending Dijon (as French) against HRE attack. They had 4 ballistae. First up they blew down my gate (which their spy had already opened), then knocked a hole in the wall next to the gate. So far, so good.

Then 3 ballistae went walkabouts near to my wall to attack I don't know what, whilst the fourth marched straight for the hole in the wall (not I suspect to make a withdrawal) where it was butchered by the defending spearmen.

Is there any way to make the AI artillery do the sensible thing and sit still and attack the nearest towers /walls and not wander off and get killed?

[khaos83_2000]

Ok, this bug will give one of your general a -1 loyalty.

How it happen to me:

I was playing in a campaign. End turn, blah blah for AI turn. Then my turn again. A general for adoption or for my little darling, cant remember which 1. I accept. Everthing is fine. Then i did some thing wrong. So i reload the autosave file, which will bring me back to the start of my turn. I load it, the adoption or proposal appear again. I accept again. Then i got a message saying one of my general got -1 loyalty trait.

[FactionHeir]

Ok, this bug will give one of your general a -1 loyalty.

How it happen to me:

I was playing in a campaign. End turn, blah blah for AI turn. Then my turn again. A general for adoption or for my little darling, cant remember which 1. I accept. Everthing is fine. Then i did some thing wrong. So i reload the autosave file, which will bring me back to the start of my turn. I load it, the adoption or proposal appear again. I accept again. Then i got a message saying one of my general got -1 loyalty trait.

This is not a bug but a chance based trigger for traits.

[redpatch]

This is the most recent buglist right?

1.) Some generals don't receive a death voice over, the camera focuses on their death without speech.

2.) Broken city gates from a battering ram focuses above the gate

3.) Upgraded towers for cities that shoot ballista bolts shoot from the towers next to gateways, but the next tower over only shoots from one side of the tower (or I'm missing the animation)

[redpatch]

Cannon/Ballista towers still don't work properly and shoot the wrong ammo.

I haven't seen this bug in 1.2

Also, I have another.

1.) When you place a grayed-out unit (say merchant) in the queue because you have no money, you can add an additional unit the next turn. Ultimately, this means one can produce up to through or four on one turn. Normally these units are limited to one per turn.

[FactionHeir]

As far as I have heard and tested, the ballista/cannontower bug is no longer existent. Are you playing the leaked version or a not properly patched version, which happens so often for 1.02? Because all other bugs you report also happen in 1.1 for instance (and have not been reported yet, although I had seen them before)

[FactionHeir]

Graphical glitch for mercenary arquebusier arm textures. It looks like they got an outstanding black frame being carried on the elbow for one type of texture (there are 2 textures for the elbow in total I think)

[DVX BELLORVM]

The battle map is broken in the area southwest of Jerusalem.



The troops are placed on a mountain cliff and you can't move them anywhere because the surrounding terrain is impassable.

It looks very interesting...

[FactionHeir]

While personally I'd say that's a bug, CA's stance on that is that it's intended. Don't ask me why...

[redpatch]

As far as I have heard and tested, the ballista/cannontower bug is no longer existent. Are you playing the leaked version or a not properly patched version, which happens so often for 1.02? Because all other bugs you report also happen in 1.1 for instance (and have not been reported yet, although I had seen them before)

I don't know if you were responding to me or the original post, but yeah, I was agreeing with you. I don't know why it's listed. All the bugs I posted I found in the most recent 1.2

[FactionHeir]

I don't know if you were responding to me or the original post, but yeah, I was agreeing with you. I don't know why it's listed. All the bugs I posted I found in the most recent 1.2

Ah ok. I didn't quite understand your post the first time as you didn't use the quote function to highlight that you were saying it was fixed, as I was saying.

[sirnoob]

This is hardly a bug rather an annoyance but when you right click on a priest that has a movement plan for the next turn it deletes it.

[Skott]

Possible bug in the random sense. Game locks up (not the pc though). English sieging HRE at Florence. After the battle is won game locks up at the strategic map where it should show winning army moving into city. So far only happened once.

[Brutal DLX]

Definitely, the siege AI has gotten worse, not better, as others mentioned already. Glaring weaknesses are shown, in my campaigns under 1.02 at least, especially with the Timurids' armies, in particular with the stacks that have elephants but no bombards. The musketeer eles march right to the gate as if they want to start ramming (as in RTW) and then hold right there and get shot up.
The ele artillery sometimes do not shoot at the walls at all, in particular I noticed this when I had no troops manning the walls (it seems as if they target the troops on the wall rather than the respective wall section).
Also, while they build lots of rams, the improvement is that they now are actually manning most of them right away, however, they stay back and wait if the first ram is successful in breaking open the gate. If it is not, more often than not, they will get movement orders but appear to be kind of stuck in place, especially when there are also siege towers placed close to them. Perhaps they equipment is placed to closely and blocking everything. This results in a complete deadlock where you either have to run down the clock or try to attack the stuck units to force the AI to issue new orders. This is very bad and can in no way be described as improved siege behaviour.
Another related problem is multi-stack sieges. It appears that also stacks that are not actually besieging (but are in an adjacent square) join the AI siege attack. This results in them showing up on the map (usually on the opposite side of the city/castle) and then staying there until the other stack has successfully opened a way into the city. IIRC, in RTW reinforcements could only join a city attack if they were actually laying siege too, this resulted in them having at least some siege equipment and thus giving the defender a harder time, as he had to deal with attacks on different sides of the town.
So, this must be fixed, as it is, I am sorry to be saying this, a real shame that such elementary things are still not taken care of after such a long time and two patches.

EDIT: I forgot to mention that attacking 2nd and 3rd level walls of castles still works only randomly. The AI still stumbles around aimlessly or, worse, lines up their main force in front of my 2nd line towers and missile troops while their troops carrying the ladders/rams arrive like 5 minutes later (if at all).

[FactionHeir]

OK, this seems to be a major hardcode bug afterall.

Remember a page back that I said how starting with the Drunken Heathen trait seemed to give a permanent boost to command even after you remove all traits and ancillaries (using plain remove_ancillary/remove_trait without operators)?

Well, it seems piety is also affected and there seems to be a trend: If your general is offered for adoption or coming of age and has traits that give a total negative value for either command or piety (not chiv/dread, but not sure about loyalty), he gets a permanent +2 bonus onto that trait.

This second piece of evidence came as one of the PBM chars came of age with 2 levels in the PublicAtheism trait (-4 piety) (and ReligionStarter 1 (+3 piety)). Well, after removing everything he still had 2 piety from nowhere.
I should note he also obtained a tutor ancillary (+1 piety), so it could be that its not only if the Attribute value total is negative, but also exactly 0, depending on if ancillaries are giving before or after traits are.

[locked_thread]

edit

[FactionHeir]

Update to 1.2 and then test. Don't test in 1.0 or 1.1 and report :p
There are some things that 1.02 didn't list as fixed but actually did fix. Not a lot but some.

[apprentice]

Family member has a Drink trait with a level of Gets Merry.
On a new turn when he gets a baby, the Drink Trait become Alcoholic.

[FactionHeir]

Nothing wrong with that really. Normally he'd get a lot of Sobriety traits for becoming a father, but there is also a 1% chance of getting more drink. Further, if you parked him somewhere, especially inside a city with the tavern line of buildings, you can become an alcoholic very quickly.

[khaos83_2000]

This is not a bug but a chance based trigger for traits.

in my case, it happens EVERYTIME.

[FactionHeir]

in my case, it happens EVERYTIME.

Then you have a streak of bad luck. The chance for that to occur is 50%. If you reload an autosave to test this though, you'll find it happens everytime because the seed is not reset.

[apprentice]

Family member has a Drink trait with a level of Gets Merry.
On a new turn when he gets a baby, the Drink Trait become Alcoholic.

It seems that Traits with nogoingbacklevels will jump to the nogoingbacklevel, when given an appropriate AntiTrait value.
For example, Trait Drink with Social Drinker value becomes a Alcoholic when a Sobriety 1 is triggered.
When that happens, Command is not reduced by 3.

Another test using the Trait Feck with a Strong Language value becomes Foul Mouthed (i.e. the nogoingbacklevel for Feck) when a Prim 1 is triggered.
When nogoingbacklevels for Feck and Drink was edited out of the traits file, the Feck and Drink were decreased using Prim and Sobriety.

Disappointing.

[FactionHeir]

It seems that Traits with nogoingbacklevels will jump to the nogoingbacklevel, when given an appropriate AntiTrait value.
For example, Trait Drink with Social Drinker value becomes a Alcoholic when a Sobriety 1 is triggered.
When that happens, Command is not reduced by 3.

Another test using the Trait Feck with a Strong Language value becomes Foul Mouthed (i.e. the nogoingbacklevel for Feck) when a Prim 1 is triggered.
When nogoingbacklevels for Feck and Drink was edited out of the traits file, the Feck and Drink were decreased using Prim and Sobriety.

Disappointing.

Tested with Drink/Sobriety setup and confirmed. [give_trait this Drink ; followed by a trigger on CharacterTurnEnd that gives Sobriety 1 at 100 if Drink = 1 -> result: Alcoholic]
Looks like a very major bug.
Time to make some fixes to the files....

[madalchemist]

(Probably someone already wrote it)

-When you win a siege battle, you have the option to ransom/release/execute prisoners, but the choice has no effect since they all die without your reputation going down; sort of an exploit, you always free them and get chivalry and they die anyway.

@Redpatch

1.) Some generals don't receive a death voice over, the camera focuses on their death without speech.

I noticed that too.

-The description for the general's character trait "Intelligent" is the one of the princess' "Intelligent_Lady" (it says the general gains +1 Charm); it is faulty only in the description, since I verified the general gains +1 command, +5% taxes and +5% trade as for the trait instead of +1 Charm.

-With France, when your spies open enemy gates, you hear your general say YOUR gate has been opened; it happens sometimes.

-Diplomatic messages (both when the other refuses or accept) from many factions say "England" almost always, even when you are dealing with HRE, Poland, Milan, etc.; for instance: "England could spare you from its plan of conquest if you give us x florins" is a message used even when it's not England who speaks.

Anyway, I play the Italian version so maybe these bugs are only in it.

[HoreTore]

@ madalchemist, I've never noticed those in the english version, so I guess that's a translation problem in the italian version.

Can't comment on the general's voice-over on their death though, I removed cutscenes a long time ago...

[madalchemist]

@ madalchemist, I've never noticed those in the english version, so I guess that's a translation problem in the italian version.

Yup, it must be a translation error.

By the way, you didn't notice the siege battle win bug or only the last 3 bugs I wrote?

[FactionHeir]

madalchemist:
Siege ransom bug I reported a while ago I think in 1.1 times already.
Death voiceover missing happens in the English version too.

Announcement error, trait descript and diplomacy are local italian version problems as I do not have them. There are messy voices for France, HRE and Islamic factions though even in the English version, but they are slightly different. Most of those I found back in 1.1 and as soon as I get more time playing those factions again I'll re-report them for 1.02 too.

[madalchemist]

@FactionHeir

Thank you for the quick answer

[Foz]

Tested with Drink/Sobriety setup and confirmed. [give_trait this Drink ; followed by a trigger on CharacterTurnEnd that gives Sobriety 1 at 100 if Drink = 1 -> result: Alcoholic]
Looks like a very major bug.
Time to make some fixes to the files....

So how should we fix this for the moment? Simply strip all the no going back levels out of the traits file? I imagine that might be best, since clearly it's better to allow traits to cancel their antitraits correctly all the time as opposed to none of the time. If everyone agrees that's the best temporary fix, I can go modify one of my existing fixes to strip those lines out of the EDCT - in fact my antitrait fix from earlier was similarly stripping the antitrait lines out of the file, so this should be VERY quick for me to do, and amazingly easier than editing the file by hand. I'd then roll out a patched vanilla file as well as a patcher to handle any EDCT at all, esp useful for modded ones... but I at least want some feedback before I go do that in case anyone has thought of something better or reasons not to do that.

[FactionHeir]

I did that with my fixes after I posted the above. (Version 1.25 to Version 1.25a)
Was actually quite easy to do. Open EDCT in word, replace all entries of NoGoingBackLevel with ; NoGoingBackLevel