View Full Version : Terrible pathfinding in towns
Ignopotens
05-24-2007, 15:22
Just something that always bugged me about RTW in general. What is the deal with the absurdly bad unit pathfinding inside towns? Why do the groups of buildings have invisible barriers extending outside their rings of walls? Why are my troops trying to march all the way around the block when there's a 100 yard gap between the lone hut, and the next group of buildings? :wall:
Especially annoying when fighting for town centers, the units don't fully engage, just sending 1 or 2 men to fight while the rest mill around, or even running out of the city center and then running back in when you order them to attack a different unit they're 10 feet from. This is most evident in towns where the center is bracketed on 2 sides by buildings. :thumbsdown:
Lastly, is there anything that can be done about this?
*edit: Also, I've never seen soldiers who were so good at running in only right angles around street corners, even if the street is just a dirt path and there aren't any buildings in the way.
I of the Storm
05-24-2007, 15:36
AFAIK, this is a hardcoded issue and there is nothing that can be done about it. It's always quite malfunctioning, no matter if it's vanilla RTW or a mod.
Yeah I hate the pathfinding.
Attacking cities is actually quite fun, but the pathfinding too often get on my nerves. :/ Especially phalanxs sucks in towns.
I've seen bugs and bad pathfinding before, but I really have to wonder what would make CA program pathfinding *that* strange. I've seen units make moves in the town square that aren't even *close* to a straight line. I'd think it would take some really strange programming to make them ignore their normal tendency to move in a relatively straight line.
I'm somewhat confident that they place "navigation nodes" at roads throughout the city to make sure they'll walk around buildings rather than bang into the walls, and reduce time spent for calculating their path. When the unit or the destination is inside the walls when the order is given, traverse along all the different node routes and compare their length to find the shortest one. Then let the unit go from node to node along the path until they have no nodes between themselves and the destination. Finally move in a straight line.
You can see this especially clearly when entering the town square. They will walk along the town square, PAST the natural point of entry to the node it decides is closest to the destination, then turn around and walk in.
Ignopotens
05-24-2007, 22:55
that seems most likely the case, it just makes me wonder how they just let that kind of stuff slide in testing (having been a tester myself for a retail game, I know someone must've brought it up), how'd that thought process go?
"Hey man what's with the pathfinding when they fight in cities, are there traffic police or something?"
"Dude, don't you know all those reports of pillaging are BS? The troops always followed all road signs, even to the point of death."
I suspect more like:
"Hey man what's with the pathfinding when they fight in cities, are there traffic police or something?"
"Dude, how serious would you say this is? Would you return the game because of this issue, or would it keep you from buying future games?"
"Man, I've forgot what we were talking about, that battle was probably like 10 minutes ago, I must have had 8 other ones since. I can rout the enemy in 20 seconds with my 20 cavalry units! I roxors!"
"Dude, awesome, glad to hear you like it."
Ignopotens
05-24-2007, 23:11
these people need some better testers then :no:
Michaelis
05-24-2007, 23:41
Couldn't agree more, I was almost yelling with frustration when I told my idiots to finish off the defenders in the town square and they actually walked AWAY, right in the other direction.
I suspect that broader settlement roads/more space around the square centre would help, but it's probably a lot of work - had a look at the file and there were all those mind-numbing coordinates.
RTW was never beta-tested... I'm convinced of it. A bit like M2TW... How can such simply bugs as units not doing what you tell them to do, which is, i dunno... Perhaps the most important function of a battle?
Or, in M2's case... Such bugs as your most powerful units on the battlefield, cavalry, not charging... And Billmen not attacking at all...
If these games were beta-tested before release, then i'm Bill Clinton. And i did NOT have sexual relations with that woman.
Those pathfinding problems a very bad for gameplay,but on the other hand they provide us with a kind of battle chaos.Something must go wrong:like misunderstood orders etc.
The real problem is that it creates terrible situations for AI units!Sometimes they perform so stupid!The most annoying thing is that they start running in the streets with no real purpose,especially when under human fire.I even prefer not to use missiles against such enemy units:)
RTW was never beta-tested... I'm convinced of it. A bit like M2TW... How can such simply bugs as units not doing what you tell them to do, which is, i dunno... Perhaps the most important function of a battle?
Actually, I know at least two people who participated in beta-testing R:TW, both notable community members, and the R:TW credits list several others. However, I do get the impression that CA was not very interested in fixing playability bugs. They probably focused on fixing the crashes, as R:TW was very stable, even on release.
What i like the most is when you order your phalanx unit to attack somebody in the city: they engage, and then suddenly turn to the other direction so the enemy can hack away at their defenseless side :furious3:
Captain: ''Attack that general unit!''
Phalangite:''But that wall over there seems much more dangerous''
Captain: ''I see... Ok, screw the general with his 150 bodyguards, EVERYBODY CHARGE THAT WALL!!!''
pezhetairoi
05-25-2007, 12:58
argh, yes! But what I hate more is the way a unit i order to move along the edge of the square to the other corner for fields-of-fire purposes will run -onto- the square. CA, wake up!
Actually, I know at least two people who participated in beta-testing R:TW, both notable community members, and the R:TW credits list several others. However, I do get the impression that CA was not very interested in fixing playability bugs. They probably focused on fixing the crashes, as R:TW was very stable, even on release.
That sounds reasonable, and explains a lot. It's true that the game has never been unstable, except when I had a faulty graphics card, certainly not CA's fault. I suppose they thought gameplay issues could be fixed later, when they saw what the most people were reacting to. I find it odd that some of the items discussed here have not been fixed throughout the patches, though. There are workarounds to all the problems of course, and they're not gamekillers, but still very noticeable.
cunctator
05-25-2007, 18:44
Somebody with 3DSM could try to mod the Data\settlement_plans\street_plans\pathfinding_XXX_XX_city.cas files. They probably contain the information what areas of cities are accessible to units and what not.
blacksnail
05-25-2007, 18:46
RTW was never beta-tested... I'm convinced of it.
You may be convinced, but you are nevertheless wrong.
I find it odd that some of the items discussed here have not been fixed throughout the patches, though. There are workarounds to all the problems of course, and they're not gamekillers, but still very noticeable.
It's often suggested that many issues were not present in the original game, but have been introduced with patches. I know this is true for the memory leak, but I don't know it has been proven for any of the other bugs. Also, what people often forget is that many of the bugs were only discovered after extensive modding and playtesting by the community. However, I do wonder how such things as the save-load bug could have escaped the programmers' notice. Don't they test their own code?
Kralizec
05-26-2007, 14:19
Well, in 1.5 (or was it already 1.3?) units occasionally spread out across hundreds of metres for no apparant reason. Sometimes when a unit is fighting, you see men moonwalking backwards while pretending they're being hit by invisible weapons. Add some pathfinding issues in settlements and some other bugs that were there since 1.0, and 1.5/6 is still by far the best functioning version of RTW IMO, wich is saying something.
However:
I suppose they thought gameplay issues could be fixed later, when they saw what the most people were reacting to. I find it odd that some of the items discussed here have not been fixed throughout the patches, though. There are workarounds to all the problems of course, and they're not gamekillers, but still very noticeable.
For stat/cost ballancing this approach is okay, but it's not limited to that. When I purchase a game I don't think I should have to wait months for a patch that will remove the siege bug, save/reload, horse archer bug...and so on.
Looking at some of the posts in the MTW2 forum it seems to be even worse for that - completely passive AI, 2-handed weapons not attacking at all due to an animation bug, and so on...in a finished product?
EDIT: fixed quote tags
Krazilec, that quote is from bovi, not me.
As Blacksnail put it in another thread (https://forums.totalwar.org/vb/showthread.php?t=84955) concerning the amount of bugs in shipped products:
This was the most recent game I could find (http://www.boardgamegeek.com/image/113720) without bugs.
People are perfectly willing to buy bugged games to get them early. Therefore, publishers are pushing developers to release early. If everyone returned every game with bugs there would be no games with bugs. Then again, there would be no games.
If everyone returned every game with bugs there would be no games with bugs. Then again, there would be no games.
that doesn't make any kind of sense. Are you telling me a professional developer team who work on the game every day are not able to fix even the major bugs? Then i wonder how the hell they could make the game in the first place :laugh4:
Yeah, sieges might be nightmarish in RTW, but i at least expected them to be fixed in M2TW. Then i found out that they are exactly as bad in M2, and the game also slows to a crawl. My humble suggestion is to stop making expansions and FIX the f*****g game first, but what do i know :whip:
Kralizec
05-26-2007, 17:04
Small bugs are inevitable and I can accept that. They're not the point, though. Stuff like the passive AI bug and the 2-handed weapon animation bug are the marks of a beta version, and beta versions should not be shipped out to stores.
People are perfectly willing to buy bugged games to get them early. Therefore, publishers are pushing developers to release early. If everyone returned every game with bugs there would be no games with bugs. Then again, there would be no games.
There's the problem. RTW was a learning experience for me. I didn't get RTW until the 1.2 patch was out for a pretty long while, but 1.3 & 1.5 were a long time away. To be fair I wouldn't return a game for the load/save bug alone but I wouldn't be willing to pay a full 50 bucks for it either (wich I didn't)
If people had higher overall standards about what they expect for their money, developers would gladly spend another 3 months hunting down the largest of bugs, knowing that it would earn them more money.
The sense is that if everyone returned games that had a bug, no developer would recover their costs. Which are legion. Of course they are able to fix the bugs. But it costs a lot. Cost-effectiveness is the watchword here.
neoiq5719
05-26-2007, 22:11
The sense is that if everyone returned games that had a bug, no developer would recover their costs. Which are legion. Of course they are able to fix the bugs. But it costs a lot. Cost-effectiveness is the watchword here.
And doesnt it cost any money to start working on patches to fix screw-ups too?
So why didnt they just do it before the release of the game and not after, u sound like by the time CA finished the game, they didnt have a single pound left in the bank.Come on is just a matter of serious work, planning and above all proffessionalism for me.
neoiq5719
05-26-2007, 22:24
argh, yes! But what I hate more is the way a unit i order to move along the edge of the square to the other corner for fields-of-fire purposes will run -onto- the square. CA, wake up!
The best is when u get into a settlement and get into a fight and with a second unit u wanna get thru by taking the highway between the wall and the fight to attack either from behind or the next unit over BUT NOOO. They just have to get RIGHT thru the middle of the battle. Maybe they are just checking up on their buddies or something .
There should be an audio file for those in which u hear. "coming thru, sorry , sorry, would u stop that for a sec" cos at least we could laugh a bit and not get pissed off.
Watchman
05-26-2007, 22:37
And doesnt it cost any money to start working on patches to fix screw-ups too?Maybe, but at that point the game's already in stores and making money right ? The longer the pre-release developement takes the longer the firm is just losing money, after all.
And you must be aware that it is most often the publishers who push for early release. They are the ones with the purse strings, and often this means that details in later development (including beta-testing) are missed.
And if you don't think that development studios can go under after successful games and with big publishing houses supporting them, take LGS as an example.
Foot
The Celt
05-26-2007, 23:41
Somebody with 3DSM could try to mod the Data\settlement_plans\street_plans\pathfinding_XXX_XX_city.cas files. They probably contain the information what areas of cities are accessible to units and what not.
Signifer1 was working on something like that. It was basically a remodeled Rome with a realistic layout and everything. unfortunately I don't think he ever got it finished. But it proves that settlements can be edited. Not sure if the actual nodes will show up in 3DSMax but it's worth a shot I suppose.
Maybe, but at that point the game's already in stores and making money right ? The longer the pre-release developement takes the longer the firm is just losing money, after all.
no they aren't ''losing'' money. They simply start making money a bit later, and with no major bugs they would likely make more in overall, no?
I'd like to think i'm not the only one who'd rather wait a few more months for the complete game than have ridiculous bugs in them, like RTW/M2TW or Gothic 3.
Sadly, the publishers usually don't agree, expecting it to be a disaster to release the game somewhat later. There are exceptions of course, where you see the deadline pushed back quite a bit (Resi 5, Alone in the Dark 5). But then, naturally, some kids come to various forums and bash to company for fixing the bugs... :no:
no they aren't ''losing'' money. They simply start making money a bit later, and with no major bugs they would likely make more in overall, no?
However, a release will set hundreds of unpaid bughunters to work, allowing you to find bugs faster. Remember, the major bugs (apart from the bad A.I., and I think this may have been a design decision rather than a bug) weren't found until several months after release.
I'd like to think i'm not the only one who'd rather wait a few more months for the complete game than have ridiculous bugs in them, like RTW/M2TW or Gothic 3.
I quite agree, but R:TW had been in development for a long time and release had been delayed more than once if my memory serves me right. At one point the publisher is going to demand that they release it and start working on the next game.
Watchman
05-27-2007, 13:03
The longer the product sits in the pre-publish stage the longer it just accumulates running costs like personnel wages, without any returns of investement. Roughly comparable to goods sitting in the damn warehouse instead of the store shelf. Of course the investors would like it to start earning the money put into it as soon as possible, even if the developers themselves would prefer to iron out the last glitches a while longer.
And what Ludens said about freebie bug-spotting.
However, a release will set hundreds of unpaid bughunters to work, allowing you to find bugs faster.
Then why didn't the bugs get fixed?
Remember, the major bugs (apart from the bad A.I., and I think this may have been a design decision rather than a bug) weren't found until several months after release.
Then what the hell were the testers doing if the (VERY noticeable) bugs weren't discovered before several months?
I quite agree, but R:TW had been in development for a long time and release had been delayed more than once if my memory serves me right. At one point the publisher is going to demand that they release it and start working on the next game.
Then i don't know, did CA deliberately leave the bugs in?
neoiq5719
05-27-2007, 17:13
u know, talking about CA´s job here is like talking to a wall. I dont know what is behind all this, could be love or whatever it is but these people have the uncanny ability to stand by CA no matter what crap they release. No point in arguing cos they will just come up with some B.S. excuse to justify CA´s horrible way of working. Period.
u know, talking about CA´s job here is like talking to a wall. I dont know what is behind all this, could be love or whatever it is but these people have the uncanny ability to stand by CA no matter what crap they release. No point in arguing cos they will just come up with some B.S. excuse to justify CA´s horrible way of working. Period.
Your behaviour on this thread is certainly not very constructive itself. Unless someone can find a reasonable way to bring this thread back on topic, whether there is a fix for pathfinding in town, we are likely to close this thread. CA bashing is neither welcomed nor wanted, reasonable criticism is fair enough, but the attacks made here are unfounded, baseless and show dodgy reasoning and a severe lack of understanding about basic business economics and business models for game publishing firms and developing houses. Basically go find proper information on this subject or stop talking about it.
Foot
Tellos Athenaios
05-27-2007, 20:19
Didn't bother to read the entire thread :whip:, but anyway here's my :2cents::
1) Bugs get found when the software gets maltreated. Testers do a lot of things, but generally have enough knowledge and decency not to do that. Casual consumers & buggers have not. ~;)
2) The CA teams are not in the position say "here's the deal, we iron everything out and whatever it takes... so be it: quit nagging us!" because they'd be outta work the instant they say it.
3) On to pathfinding: well, based on my experience with RTW...
a) There appears to be a certain center within a unit, which is used to determine the general direction a unit has to go to... (forward, turning etc.);
b) However especially with large units: they don't fit in the smaller streets, or on walls, or in siege towers or between trees -> some of the unit gets stuck making the general calculations based on the unit center pretty worthless since the center isn't anymore where the unit itself is;
c) Buildings & trees count as obstacles (obviously), but hills on which buildings are built only partially count as such (that's why those ridiculous fences and stuff is place around them... makes it slightly less odd);
d) Those public gardens etc. count as obstacles as well, however a tree standing pressumably on the other side of the road, in engine terms is in fact part of the obstacle. May have been done with the idea "the garden is always to be surrounded by other buildings & this saves PC performance" in mind, may be not... whatever that's the case;
e) Units which are grouped are programmed to stay in formation; and to do that there's this "group center" simply the idea behind the unit center is being applied to multiple units as one unit of it's own.
Hopes this clears things up. At least thread's :focus: :yes:
blacksnail
05-29-2007, 21:59
u know, talking about CA´s job here is like talking to a wall. I dont know what is behind all this, could be love or whatever it is but these people have the uncanny ability to stand by CA no matter what crap they release. No point in arguing cos they will just come up with some B.S. excuse to justify CA´s horrible way of working. Period.
Railing against the game companies that have betrayed you is a cathartic pasttime, I'm sure, but it gets really old, really fast for everybody but you. I can't speak for the rest of the team, but I get annoyed when people whine about stuff that can't be changed when they are well aware that it can't be changed. There are things in this world which deserve your ire far, far more than a computer game that doesn't do exactly what you want while interacting with hardware and software that you don't understand.
I know, what irony; an EB developer says this to anybody. That should underscore how obnoxious it can get, and it can get really obnoxious. There was some guy who threw a tantrum when we told him the siege tower pathing was hardcoded and blamed us for not fixing it. What can we possibly say to someone like that without coming across as insulting? "I hope that at some point you learn to enjoy things?" I really have no idea.
Ignopotens
05-29-2007, 22:10
hah I know exactly what you mean
I didn't realize the pathfinding was unchangeable when I made this thread, or I wouldn't have made it at all
blacksnail
05-29-2007, 22:37
See, that's fine - if you don't know about the obscure arcana of the RTW engine, I encourage you to ask about the obscure arcana of the RTW engine. ~:) This isn't something that you should know or care about unless you're working on a mod.
vBulletin® v3.7.1, Copyright ©2000-2025, Jelsoft Enterprises Ltd.