1.2/1.3 Community Buglist

[frogbeastegg]

So the sudden and drastic change in AI behaviour and the fact Scotland's AI woke up doesn't remind anyone else of a reverse version of RTW's save/load bug? Because it strikes me that way - that's why I'm rather desperate for others to take a look.

It all started to happen on the very turn I reloaded my game. I didn't take more territory, I didn't declare war on anyone, I didn't suddenly raise a massive army, I didn't do anything shady with agents - I can't explain the sudden massive increase in aggression, or the fact Scotland sat about unmoving for 50 (took 2 turns to move an army to York, then went dead) and sprang back into life the very turn I loaded my game.

[FactionHeir]

Well I think there is something like that frogbeastegg (I think I mispelled this twice before getting it right just now).
Reloading a game (usually to fix the train/retrain issue) sometimes resulted the AI declaring wars on each other the next turn, but mainly due to an actual updating from the campaign_db_ai file it seems.
I.e. trusted allied will auto declare war on their ally's enemies and this usually happens but not always. Reloading/end turn often fixes that so thats probably what you were encountering (not the ally thing, but that file controls all AI strat map behaviors)

[sapi]

So the sudden and drastic change in AI behaviour and the fact Scotland's AI woke up doesn't remind anyone else of a reverse version of RTW's save/load bug? Because it strikes me that way - that's why I'm rather desperate for others to take a look.

It all started to happen on the very turn I reloaded my game. I didn't take more territory, I didn't declare war on anyone, I didn't suddenly raise a massive army, I didn't do anything shady with agents - I can't explain the sudden massive increase in aggression, or the fact Scotland sat about unmoving for 50 (took 2 turns to move an army to York, then went dead) and sprang back into life the very turn I loaded my game.

That sounds horrible

I'd check it myself, but my 1.2 install has somehow picked up the ctd bug in sieges

[Foz]

That sounds horrible

I'd check it myself, but my 1.2 install has somehow picked up the ctd bug in sieges

That's not permanent, at least wasn't for me. If you wipe, reinstall, then patch to 1.2 again, it should hopefully work.

[wilddog]

I think the main singular issue is still around the passive AI for me. I just played a game and came out to sortie. The AI just stayed in formation as it normally does allowing me to shoot it in the back etc as before. It had however appeared also with siege bits (ram and ladders) on the 1st go of the siege (didn't think it was suppose to build them so quick) and then refused to put them down to fight as I walked around them!

On the plus side the earlier crash I had with sieges was me being stupid. I had copied over my older CFG file and the new cfg had more attributes generated than the old one so I'm guessing that's why I then had problems.

[Gaius Terentius Varro]

I think the main singular issue is still around the passive AI for me. I just played a game and came out to sortie. The AI just stayed in formation as it normally does allowing me to shoot it in the back etc as before. It had however appeared also with siege bits (ram and ladders) on the 1st go of the siege (didn't think it was suppose to build them so quick) and then refused to put them down to fight as I walked around them!

On the plus side the earlier crash I had with sieges was me being stupid. I had copied over my older CFG file and the new cfg had more attributes generated than the old one so I'm guessing that's why I then had problems.

https://forums.totalwar.org/vb/showp...76&postcount=1

[wilddog]

Gaius Terentius Varro - Sorry don't understand your post? Are you saying that the item I posted which is happening ALL the time as far as I can tell is the intermittent AI problem that CA mentioned? It just doesn't fit normal definitions of intermittent.

On the other hand I just experienced this one. Also in a siege situation. I'm hoping this IS intermittent.




Depending on angles you could also get to a lone figure about 4 metres in the air (It crashed after I tied to get the snap shots).

[Daveybaby]

Dali mod?

[FactionHeir]

Dali mod?

lol, priceless.

Personally I was thinking of a mutated ballista

[Foz]

Just wanted to lend some further explanation to the problematic pathfinding for units chasing routers. I've done quite a lot of router chasing tonight, and have noticed something: the problem always occurs when the unit being chased has become spread out for some reason. Often a few of its men failed to break away from combat right away, or got hung up on the city gate while fleeing, but whatever the reason, men being separate from the main clump of the unit is the cause of the trouble. Every time the unit being chased had good grouping, the chasers rode right through the middle of it and wiped them out.

Noting this behavior, I began holding the shift key down when I had a unit selected that was chasing routers. For those of you not familiar with this, it displays a red pin on the map that shows you exactly what point your unit is heading for. I noticed that units that were spread out typically had a pin nowhere near any of the men in the group: one man far away often made the pin appear in the space between him and the rest of the unit, which was always where my horsemen would ride to. From this I deduce that the game uses the "center" of the unit for all pathfinding in combat: basically, the average position of the men in the unit is considered to be the point the unit is at.

The problem, then, has nothing to do with pathfinding or unit cohesion of the chasing unit, but rather has to do with the cohesion of the fleeing unit. Because routers are allowed to flee in complete chaos with no formation at all, the center of the unit (as calculated by the game) goes wild, and often leads the chasers to a place where no actual men of the unit are. Needless to say, this looks completely dumb, and has been the subject of much stress to some Orgers.

What then can be done about it? The first (though likely least nice) solution is to simply enforce some sort of ordered formation on routers. This would be totally effective, however it would kinda lose the effect of routers running for their lives in, well, disorder. Perhaps semi-orderly formations could be used for routers, to preserve the feel while making sure not to leave those problematic stragglers behind.

Likely better is to change the formula for finding the center of a unit. The problem would likely drop off the radar if some substantial portion of outliers would be ignored when considering the center of the unit: that is, if we're ignoring 25% as outliers, the 1/4 of the unit that is the most average distance from other soldiers would no longer count toward finding the center. This would essentially treat the clumped men as a unit and ignore the few that are quite far away, meaning the chasers would hunt down the most significant portion of the group they're chasing, and do so totally effectively. There can still be trouble, though, if the rogue portion of the group is more than the chosen outlier percentage: they would again tug on the center.

Another possibility would be to set a threshold distance a man can be from the others to be considered. Using the median distance from a given man to every other, it should be simple to lop off any above a reasonable level, and thus arrive at a much tighter calculated unit center for pathfinding purposes. I recommended using the median because it neatly classifies a man as part of the bunch or one of the rogues, since his median distance from other men will be high if most are far from him, and low if most are nearby. So since the median will be much more polar than an average, it is IMO the best choice for dealing with this problem, and should result in very good unit center calculations that are not influenced at all by nonconformists.

As far as I can see, removing outliers by either method would not substantially affect anything in normal circumstances: units in good order should generally continue to have their center calculated very near where it was the old way, while either should show significant benefit in chasing routers. I of course don't know the particulars of the situation... but I do hope my comments on the matter are some use, provided the devs haven't already fixed it (a change could be in the new 1.2 build already for all I know).

[PutCashIn]

I was wondering about the chasing issue, thinking back to RTW I remember the units would move untill the enemies standard was in the centre of the chasing unit.

Now in M2TW it looks like the chaser is moving unitll the front of the unit coincides with the emeny standard, then slows down to do some chopping/hacking, but at least half the enemy unit still gets away.

Basically in RTW, the 'footprint' of the chaser (seemed to) coincide with the footprint of the routing unit.
Now only the front half of the chaser coincides with the back half of the router.

If any of that made sense....?

[Gaius Terentius Varro]

Gaius Terentius Varro - Sorry don't understand your post? Are you saying that the item I posted which is happening ALL the time as far as I can tell is the intermittent AI problem that CA mentioned? It just doesn't fit normal definitions of intermittent.

On the other hand I just experienced this one. Also in a siege situation. I'm hoping this IS intermittent.




Depending on angles you could also get to a lone figure about 4 metres in the air (It crashed after I tied to get the snap shots).

No I am saying that the problem which is happening a lot to me ( not all the time tho) is on a "to do list" for the new patch which i find a relly good bit of news

[nikolai1962]

So the sudden and drastic change in AI behaviour and the fact Scotland's AI woke up doesn't remind anyone else of a reverse version of RTW's save/load bug? Because it strikes me that way - that's why I'm rather desperate for others to take a look.

The scotland/york thing is a side effect of a pathfinding error. They can snap out of it sometimes, even without reload, though i'll try that out to see if it has a consistent effect (if it does, then it's a good thing). On the other hand the rest of the things you mention sounds like VH behaviour where your faction standing has been tanking for 50 turns whereas i think you said you were on M campaign diff. So that is a bit worrying for people who like M difficulty precisely because of the lack of AI "attack the human no matter what" stuff.

[FactionHeir]

Foz: Your observations are correct. I had quite similar experiences.
There is also another problem though and that happens if the routers are very few in number or at least one of your chasers overtakes the routers. In those cases your chasers tend to stop, regroup and never get anywhere chasing them.
Further problem is if the routing unit is nearing the end of the map. Then your chasers often for some weird reason go either left or right of the routers, run against the edge of the map and then against the routers by which time they will have left the battlefield.

[Gaius Terentius Varro]

I am having a problem whith the venetian units armor upgrades . I applied the usual fix in the export_descr_unit.txt but the model upgrade stays in the same order. After the patch i mean
Edit: ok i forgot the mod switch so ignore

[econ21]

I think there may be a bug with lance cavalry charging other cavalry - there does not seem to be the charge bonus you get when charging infantry. This is from another thread:

The charge of cavalry against another cavalry unit head-on has no effect, as far as I can tell (in 1 on 1 battles at least). Nobody dies until the secondary weapons are drawn.

I think that the lances are too short so when the mounts come head to head the lances don't reach the rider(s) - and nothing happens. Since they also have the same mass, they cannot push into each other's formation either. I am not sure whether this is intentional, but it sure doesn't look like the tales of medieval jousting matches.

May be saying nobody dies on impact is too strong, but there does not seem to be a benefit from charging.

[FactionHeir]

Yeah, only if you charge cav from the rear do they actually take extra casualties and then mostly only if its a 1hp unit. Head on hardly ever does any damage, but highend cav charging low end cav head on will still kill the first row of cav if the cav being charged is static. If they both charge each other head on, noone seems to die.

[Foz]

Has anyone noticed that 1.2 cavalry ignore the alternate attack command entirely? I'm not sure if it existed in previous versions, but I've had occasion several times lately to attempt making my knights ride in with swords drawn, however they always seem to bring out the lances instead. This is especially annoying if I'm pretty sure they can't correctly charge in the current situation, but they raise the lances anyway, only to run up to the enemy unit with lances up, stop, and draw swords anyway. The usual result is much wasted time that would not happen if I could get them to ride in with swords out already...

[FactionHeir]

Happens in 1.0 and 1.1 as well. Found it annoying too. Kind of defeats the purpose to even have a charge value for secondary weapons...

[TinCow]

For the first time in my life, I think I've found a bug that I've never seen reported before. It's not serious, because it requires a lot of work to achieve, but I figured I should report it. In short, I have a family member that isn't showing up on the family tree. Here's how it happened:

I had been refusing marriage offers for my young faction leader, because I wanted to use him to arrange a political alliance with another faction. On the turn that I finally tracked down a 4 charm Portuguese princess, I was offered a marriage proposal for him yet again. Instead of rejecting it, I simply minimized the offer. I then approached the princess, she accepted the marriage offer, and he was married. The other marriage option didn't disappear though. As a laugh, I clicked Accept on the marriage offer, and was given the same message you get when you adopt someone who has already died. Interesting, but no big deal.

A few turns later, a marriage proposal screen popped up, like you would get if you had an unmarried Princess wandering around. That was strange, since I had no princesses at all. I also noticed that the name of the woman was "Campagnola de Crete" which was the name of the woman on the 'aborted' marriage proposal to my faction leader.



I minimized the offer and went to check my family tree to make sure I wasn't imagining things. Sure enough, there was no one by that name on my family tree. In fact there weren't any unmarried women (other than children) on my family tree.



I wanted to see what would happen, so I accepted the marriage proposal. It went through as normal and the new general spawned into my capital.



I then checked my family tree again, and nothing had changed. Mouseover of the new general showed him as a family member (see second screenshot), rather than a general, but he wasn't on the family tree at all.



Not a serious bug, but I find it kind of interesting. I'm now curious to see if he'll have any children, or whether he'll just act like non-family member general.

[FactionHeir]

Two more bugs:

- AI can see your hidden forces if they are sallying from a battle. In my case, they sallied from Rennes and I had my mailed knights run to a nearby forest and hide. Even though I placed my general away a fair bit and the AI was still busy regrouping his troops at 100 meters from the gate (according to the files they do this before anything else) while my knights were already hidden, he later directly charged at the hidden units, ignoring my general.

- Regicide missions do not expire when the mission giver dies. Rufus offered 5000 for William's death but I decided to kill Rufus first and see what happens, but the mission continued and I even got 5000 for killing William the next turn.

[Husar]

I took Antioch in a crusade and then the Mongols came and besieged it with a single stack, I defeated that stack in a siege battle and had only few units left. Next turn they besieged again and in the following turn they attacked with three stacks. I thought I'm gonna show them and had my longbowmen deploy their stakes in the city centre, hoping that they'd lose a whole lot of cavalry trying to take the city but when their ram broke through my gate, the game crashed. I reloaded and tried again, game crashed again at the same point. I then autoresolved. Rest of the campaign seems fine so far, haven't fought the Mongols since, Antioch was my only province there.

[tex_-]

I just installed leaked 1.2 over 1.0 and game crashes everytime i try to run m2tw. This didn't happen on 1.0 or 1.1 version, don't know what's the problem, does anyone had same problem as me? Plz help

[HoreTore]

Tex, that would be the reason why it was withdrawn.

You'll simply have to tweak around and uninstall/reinstall a bunch of times, and try to solve it.

There is a bug in the leaked patch causing a CTD in certain circumstances.

[FactionHeir]

- Allied AI spies will cause a PO drop if they infiltrate your settlements but if you infiltrate an ally's settlement, the spy does not cause a PO malus.

[Husar]

I think with the twohand fix and the shield fix now, the balancing is a bit off, I hope some two-handed units will get a significantly higher attack in the official patch.
I just had some twohanders from the HRE fight against french DFKs in an open plain custom btlle and the DFKs won with more than half of their men left. I tested this after noticing in my english campaign, that Armoured Billmen, being a higher tech level than Armoured swordsmen, are definitely inferior in melee. They may have a bonus against cavalry, but the higher protection of units with shields usually makes up for a lot, sword and shield infantry are now the kings of the infantry, don't know whether that's right.
Two hander animations are fixed, but as long as they have mediocre attack and rather low defense, they're only useful against archers and maybe early spearmen and that's it.
Twohanders have just one more point in attack(in MTW they did at least have armour piercing) and significantly lower overall defense than sword and shield units, I hope that won't be the case in the official patch anymore.

Apart from that, I like the balancing a lot, dismounted english knight, who also use two handed weapons, seem fine, low defense but a very high attack which makes up for that.
It's just badly armoured, high-tech twohander units which also have only a mediocre attack rating that are apparently useless now.

[RihardsVlasovs]

in file descr_faction_standing.txt there is trigger
0080_F_Military_Assistance

trigger is supposed to increase global reputation of faction that is PROVIDING military assistance.
But instead this trigger increases global reputation of faction that is receiving
military assistance.

It is essy to test (1)by changing trigger value to big number, and (2)then within the game gift atack faction aggreement against rebels to papal states and (3) then just attack rebels. The result: your faction reputation stays the same, while papal states reputation becomes trustworthy.

[Slicendice]

I've experienced problems chasing routers in seige battles as previously mentioned. Units sometimes will not chase fleeing enemy at all. Charging cavalry into fleeing crossbowmen makes them stop amidst the bowmen but cavalry does not engage. If left alone crossbowmen will go into "fight to the death" mode and will kill the cavalry (in this case generals bodyguard). Cavalry will kill units infront but will not kill units behind it. Also Cavalry will maintain it's tight formation and stand still. This causes mass casualties on HC if left alone.

When assembling units (spear militia and peasant archers specifically but perhaps not exclusively) inside enemy city some units assemble in loose formation. Clicking on formation button to make them loose and then again to make them tight does not always work. If it does, making units into straight line can again make some units loose formation.

Units attacking city wall with ladders and towers marched up the ladders and suddenly fell back away from the ladders even though some were on the wall already. In fact all units on ladders and towers did so simultaneously. These units were not grouped. I believe that a wall breached by my balista triggered the withdrawel, however, the gate and another wall were already breached.

Units, (mercenary crossbowmen) will not respond to command get on enemy walls to fire on enemy even though wall and entrance are nearby. Also same units may not fully get on wall if it does respond thus leaving some on the ground and some on the wall. (Running icon is present however no movement.)

Attacking peasant archers firing on defending General Body Guard in the townsquare triggered the Body Guard to move toward units that were down another street. ie. archers were south, GBG moved east toward Mailed Knights. Not sure if GBG would have moved down that street if it was empty though. Is it possible he went down there to avoid archer fire?

Experienced game slowdown in the campaign map. Also memory leaks during seige battles.

Only in my first 10 turns so no field battles yet. I'm already dreading what else I might find.

I think I might go back to leaked patch. It seemed better than official patch.

[Foz]

Attacking peasant archers firing on defending General Body Guard in the townsquare triggered the Body Guard to move toward units that were down another street. ie. archers were south, GBG moved east toward Mailed Knights. Not sure if GBG would have moved down that street if it was empty though. Is it possible he went down there to avoid archer fire?

Couple thoughts on this one. First, yes it's entirely possible he went there to avoid the archer fire. The AI may actually have opted to try to engage the mailed knights, hoping to force your archers to stop firing as you'd take friendly fire casualties. Also, I noticed the AI actually send mailed knights out of the city square to combat my mailed knights that were positioning for a charge (though some distance down a street from the square). So the AI might have decided to send the GBG after your mailed knights simply because it thought it had a shot at beating them, or if it considered them a threat worthy of its attention. I nearly fell over and died when the AI actually sent a unit out of the square after setting up camp there, as I honestly don't think it ever happened before for me, and thus surprised me that much. I can only hope this is a sign of a more proactive AI, even in a defensive posture, as it has all-too-often simply sat there and taken abuse from archers, or allowed my units to flank it unhindered.

[sapi]

Even in leaked 1.2 (i'll be continuing that campaign before installing the official patch) I've been noticing the enemy move out of the town square more, even taking all their units out and rushing towards my men as they ran for the destroyed gate in an attempt to keep me out, so hopefully it is a change for the better...