Carrier Command Beta due out tommorow...

[Drewcifer]

Originally Posted by soopaman2

I caught this on my steam window, and if not for other games I would have bitten.

I liked being able to control different units, but is the transition fluid, or do you have to choose one or the other? Like can I go from a plane, to a tank. etc?

You move from Carrier-Manta-Walrus instantly with the click of a button no matter where you have sent it no matter what it is doing and take direct control, the second you leave it the AI picks it up and continues doing whatever it was doing. Unless you have say taken something so far off course that it is unsure of what it was sent to do it will go into suspended mode and wait for updated orders.

[CaptainHaplo]

OK - for the record I was one of the first P&C members. Bought it 3 days after they opened the test program. I later upgraded. I played all three demo's and the patched version.

#1- the campaign
It is a VERY long tutorial. Overly long. The FPS portions are there for a storyline - but honestly it was a mistake to put this in. Lacking in story, character attachment and usefullness. I can not recommend the campaign other than just to learn how to do things. For that, however, it is very useful.

The only reason to play out the entire campaign is the on the final mission you fly against a recorded flight by one of the devs and you have to shoot him down. Bragging rights at best?

#2 Strategy challenge.

Very good - but it needs a little tweaking. Additionally, the ability to allow for non-aligned islands - and control over how many islands each side has to start - is very important and will give some legs to this mode. Overally, still very enjoyable.

#3 Gameplay
Wow - I have to say - I love this game because of how it plays. It is faithful to the original to a point, yet takes advantage of a lot of the improvements that were needed. Some people don't care for some design decisions - and a few stick out in my mind as well - but overall its great fun. The strategy for attacking each island - balancing the tactical and the strategic along with the logistics - is just great!

"Walrus" pathfinding..
OK - I guess I take issue with this. I am going to paraphrase what someone once said about Command and Conquer....

Its Carrier Command - not walrus micromanage. The AI is servicable - and that is all it needs to be. The idea here is these are remote drones - with limited ability without a human in control. Being upset because they don't travel well in groups - is silly. The players who complain - are wanting the walrus's to do their work for them - they want to just "assist me" and then run around shooting anything that moves, and wondering why they left their "help" back in the dust.

The game makes it clear they are limited in ability. It makes SENSE. If the players took that into consideration - stuck one AA walrus assisting them instead of another ground assualt, an AA and a "healer" walrus all together, and then went flying up ridges and down valleys where the AI is bound to struggle, they wouldn't have cause to complain.

On the rare occasion I allow the computer to control a walrus - its an AA one that is there to simply follow me and provide antiiair cover. It does great following me and knocking enemy manta's down. Only struggle is when I go up a huge rise or down an incline instead of taking the road that goes the the same spot. I sacrifice a little health for speed - the AI won't. I see no reason to complain about that.

Really - its a non-issue for those who don't want to just tell the drones to go conquer something and sit back to read a book.....

[Skybird]

Hm. I got the impression it is not that simple, Haplo.



It is not about AI drone having the war brought to victory for you. It is about being autonomous enough to be able to negotiate a path around a simple rock or tree.

Compare to the AI of NPCs in Skyrim, for example. They are able to keep up with you if they are your followers. They meet an obstacle hindering the direct line from their position to yours? A ridgeline? They immediately evade to the side, flanking the obstacle, finding a route, sometimes quite complex routes and on multiple levels, and reach you reliably, with a delay. - So much for autonomous pathfinding.

Compare to the AI in SBP. You lead a formation, and even while driving you can alter formation type and width. Your AI buddies reliably accelerate or slow down to fall into their right place. When holding a battle position, they will autonomously react to the presence of detected threats, according to their firing orders that allow them to aggressively engage, to only return fire, or to never shoot back, they will stray away slightly their formation positions to find better firing positions and protect their hull behind terrain features as best as possible. After threads disappeared, they will fall back into positions. Infantry chooses by rigidity of defend-command whether they stay outside building, seize the ground floor, or take up positions as high as possible. Small groups of obstacles every unit can negotiate by itself. Granted, the AI in SBP is not good at detecting small holes in the ground with water that could get a tank stuck, and it is more easier confused in later versions n that in early ones when it gets trapped in a street with too tight buildings. But all in all the system works quite well in 6 out of 7 situations or so. - So much for formation AI.

But how is it with vehicles in the ArmA series? Simply terrible. The absence of believable physics does not help to improve the situation. Have they ever healed the issues in any of their titles over time? No. Many patches sometimes, long support. But improving the AI? Sometimes claimed, sometimes listed in readmes, but not convincingly taking place.

Check their forums. The issues with Carrier Command's AI and pathfinding, says their forum, has been fed back to them since the Betas, since months. They did not react to that. On Monday, a representative or speaker of the developer first time showed up in their forum, sayng big and loud and warm Thankyou for the cheers as if there were no criticism at all, and opening two subsection in the existing forum, one about technical issues and bugs. Some of the already existing threads about other bugs got moved there. None of the then three existing threads on Walrus and pathfinding AI got touched and moved. He also answered in several threads to problems people had - but did not answer a single time to the intense criticism of the pathfinding AI. Yesterday or this morning, a new long thread was opened in the technical subsection, about the AI. Again, no reaction to that. And somebody else, a newcomer, asking about the AI issues, got warned not to scatter the forum. Now, over the week I did several Google searches for customer feedback and reviews. And I stumbled over comment by peopel saying that their posts in the BI forum, and whole threads, which were critical of the AI, got deleted, so they claim.

I think it is fair to say that the game has potential, but suffers from some serious bugs, and some of them are showstoppers for many people. Lousy AI affectes not only the Walrusses, but the Carrier and the Mantas have issues, too, though no such crippling issues. I think BI has a well-documented record of not knowing garbage about how to do AIs, and that it would have been good advise for them to hire somebody from outside to do that part of their job (a recommendation you hear often in present web comments). I think the chances are very thin that they will fix the AI, and I think their months-long non-reaction to reports about AI and their forum policy indicate that they intend to just ignore it and keep it hidden as best as they can to not hurt their sale numbers.

So: outlook is negative. I wish they will teach me for the better, and surprise me and all of us, but I am sceptical - very. I do not know how the AI is implemented and whether it is available for open modding, if it is, then there may be hope. But then I am not willing to pay them for letting others do their homework, and will consider buying it not before one year or so is over, the issues have been modded away, and the price is no longer two-digits.

Great potential, partially good execution - and allowing laziness and incompetence messing it all up seriously. That's the impression carrier command currently gives to many people.

I do not get it. Since almost a decade BI takes heavy flak for their incompetence to master AI issues and reliable pathfinding - only that the ArmA series is multiplayer saved it from disappearing from the scene. If it were a single player title only and not offering better AI, nobody would talk of it anymore today. But CC is single player exclusively, and thus a working AI is vital for it to live. Have they pledged an oath not to learn and not to accept assistance from an expert for AI tuning that they could hire, and not to listen to the community telling them over and over and over again?

[TheDude107]

Originally Posted by Skybird

Great potential, partially good execution - and allowing laziness and incompetence messing it all up seriously. That's the impression carrier command currently gives to many people.

I do not get it. Since almost a decade BI takes heavy flak for their incompetence to master AI issues and reliable pathfinding - only that the ArmA series is multiplayer saved it from disappearing from the scene. If it were a single player title only and not offering better AI, nobody would talk of it anymore today. But CC is single player exclusively, and thus a working AI is vital for it to live. Have they pledged an oath not to learn and not to accept assistance from an expert for AI tuning that they could hire, and not to listen to the community telling them over and over and over again?

It could be they wanted to keep the gameplay the same as the original(3 same units, same gameplay, same two modes) but knowing BI as of recent, I would assume they skimpped out and got lazy. ARMA brought them back from obscurity. ARMA 2 was a success.... but it would have been a very minor success without Dayz mod. Maybe even a failure.

And after looking at previews of ARMA 3 it seems like BI is getting a little lazy concerning their games because they decided to make fictional units for the upcoming title set in Limnos.

The "reasoning" behind this was that they wanted to create something of their own and relax a bit after ARMA 2. Which is understandable. Until you realise it takes too much money to get the rights to model some real life vehicles and weapons, which probably scares them when looking at pre-Dayz mod sales numbers(Which were sad). It probably is alot harder to accuratly model the damage done to a fictional vehicle. Even harder for the fans to claim that the damage was modeled badly because the vehicle doesn't exist in any form, nor does any stats on it outside the game. Thats just a wild guess from a long time fanboy, though.

Hope my favorite developer isn't going down the COD road due to getting cocky. They will fail hard if they loose their sim fans. We may be small in number but we are dedicated to our games and shell out cash for value. Regardless the fans will yell at BI until they release the mod tools and we will have re-skinned units+new models for years and years to replace anything lacking in quality for the real deal.

[Skybird]

I fail to see in how far their desire to stay close to the original - which I know, I played it back then on Amiga - limited them in implementing a reasonable algorithm for pathfinding, and do some better bug squashing. In fact the feedback recognises that they seem to have done a good job in designing the game - saving the principal design of the original, but improving it by using additional features. The design idea of their sis not the problem. Everybody agrees that by what it wants to be, the game indeed is deigned great. I also have that impression. Time enough they have had for sure. CC has been first announced - how many years ago? Six? Seven? Or even more? I think it was short after the Gulf War 03 that I first posted a link about Carrier Command being redone. But memory could fool me.

[Skybird]

Maybe they are serious when having set sail for proving me wrong...?!

An official comment/announcement has been released. They formally recognize the existence of "problems" with the AI and pathfinding, they recognize that these are widely known by now, and they announced to have started working of a dedicated patch for these issues, indirectly indicating that it is a huge task.

I would be happy indeed to be proven wrong, and would be the first to admit it. I recognized above that BI indeed offers longterm patching support for their ArmA series, I just criticized that in the past they were not successful in patching AI problems. If this time they score a first in achieving that, I'd be happy.

[Arclight]

Actually, they were pretty succesfull addressing issues like this in the past. Like I said before, ArmaII AI couldn't drive down a straight road initially. Nowadays I can issue a move order to an AI driver to go clear across the map and they drive there without much issue.


I gave the demo a try (btw there's a demo) and got through it without much trouble. There's definitely issues; when I had more than 1 following me #3 and 4 would end up bumping and trailing behind. One of them failed to drive across a bridge. Solution: make teams of 2. #2 follows 1 and #4 follows 3, orders given to 1 and 3.

Come to think of it, Arma can't handle convoys either. Not without hilarious results at any rate.


Think I'll give the full game a shot soon. After trying the demo and experiencing the issues first-hand I don't really see what the big deal is. Even if only due to years of experience with BIS AI and knowing how to make it work.

[CaptainHaplo]

Originally Posted by Skybird

Hm. I got the impression it is not that simple, Haplo.

Well, Arclight seems to have found the same experience as I did. One Walrus following you does fine, make a whole team of them and the last 2 have trouble. As I said - for people wanting to just run around and have the AI cover their backs, this is not its purpose.

In fact, in the campaign, there is a part where they specifically talk about the "limited AI" that the vehicles have. The units are all field adaptations that allow remote control - thus the AI itself is considered a "new" edition to the unit in the field and would be expected to have performance issues.

The examples you cite, Skybird, are excellent points. The problem is, each one is an AI that is supposed to be emulating the smarts of a human. Either a AI squadmate or a human type AI driving a vehicle. I would expect them to be somewhat "intelligent".

However, if a game even tells me they are "limited" (I even think it uses the term "crude") then I have a different expectation. Especially when the entire storyline accomodates that fact.

Having been part of P&C from nearly the outset, I have extensive testing time on all three betas. During that time, Walrus AI was never an issue for me. Mainly because my playstyle doesn't expect 3 other tanks to run around with me, one to protect me from Manta's, one looking for ground threats and one to repair me if I get damaged so that I can just rampage around and feel cool driving what would then be a nearly invulnerable tank around a pretty island.

It is all about how you use the units you have. It is Carrier Command. Back in the day, since you remember the original, there was no AI for your units - you controled each one or it just sat still. They have improved on the original in some ways. Just not to the point where the "player" can select his four "tanks" - tell them to attack the island, and then go heat up a pizza while they do his job for him. Now will the AI allow the player to create a team of AI's that will provide unbeatable support to the player. It does this by making the player make tactical decisions. Will an AA walrus be better - or will a healer walrus do the trick? That makes the players choice of weapons on his own tank critical - laser or plasma? Each decision feeds another - which shows some really great design choices.

I don't have a problem with that. I don't have a problem with an AI that isn't "smart". Its all within the scope of the game.

Let em fix some of the pathfinding issues. It would make the final missions of the game much easier (as you do have multiple walrus's following you to provide support). Great. But the complaints are well out of proportion to the needs of many players.

It really comes down more to play style and preferences. If the player wants to make it a FPS (and there are questions on whether that style of gameplay can be modded in from the campaign), then they want all the support they can get. If they are an old school player who doesn't rely on the computer to play against itself and win - then its pretty much a moot point.

The only time walrus pathfinding has been anything of an issue with me has been the rare niggle where it doesn't recognize an obstacle in the way, or when winding paths make it take a very lengthy, back and forth route. Even in those cases, it simply takes a little longer than a human driven unit would. That to me is reasonable.

[Skybird]

One really should differentiate behavior AI from pathfinding AI. The issues with the walrusses that get reported time and again, are almost all pathfinding-related.

You implied that my referring to Skyrim followers and SBP tanks acting smart, and "intelligent". Well, the movement of the enemy - is completely scripted by the mission designer, paths and triggers. Sop is the behavior of units in my formation who adapt to my own tank'S movement when I am the leader, or who follow the lead tank if it'S AI has a path-and-trigger script to follow. The followers in Skyrim only know how to not launch a fight if the enemy is not aware of their presence, or to go for it if an enemy is within reach (internal triggers that can be accessed via some mods, originally this was not open to the player). Skyrims followers are only very competent in pathfinding. It is remarkable how they find a way around even a whole mountain, to catch up with the player, with you sitting on the peak and watching them. Trees, stairs, cliffs - usually no problem, they zig-zag home to mama and usually in the shortest possible way.

If there is something like a real set of AI for autonomously engaging in , disengaging from and manouvering during combat, then it seems to me that you find it in dogfighting simulations.

The problem with CC is not that people expect walrusses and mantas to be smart like humans, but that their pathfinding has serious issues. But in a single player game where you control several units simultaneously, a pathfinding AI for mastering even the basics, is a must. You give a string of waypoints, and then you should be able to expect that they follow these waypoints, not being stopped by a single tree or a rock in their way, or not landing on the beach. When instead a walrus slams into a rock and spends the rest opf the mission with spinning around it, or when it meets a bridge and simply falls down from it or ignores its presence, then you have indications for a seriously broken pathfinding AI. When Mantas fly and the ground elevates and climbs into a mountain peak and they do not climb and do not fly around, but slam right into it, then there is some stuff seriously broken. I do not expect them to be smart like humans, nor have I ever read in a forum post that somebody else does. What I expect is that when you leave the AI with an order to move from here to there, it just does that and is able to use bridges, avoid mountains, and can turn around small obstacles like trees and rocks and can climb a small step or a shallow hill by itself. I do not complain that they get shot into pieces if I am dumb enough to give them a path that exposes them to enemy fire. I do not expect them to fight for me and conquer the base for me all by themselves. I only want them to work reliably with getting from A to C via B, and that they return fire or engage when they take fire or somebody enters their firing range.

BTW, the old classical CC had the ability to have one unit move on by itself. You saw it in a separate small window when having teleported back to the carrier, and you could equip it with a range extender for the remote control signals from the carrier. You could jump into the cockpit any time you wanted. Mostly this was used with Mantas, since walrusses were so slow even back then. There were also fuel tanker kits available, to refuel units that had run out of fuel. Pretty innovative for that time. And very difficult - I never defeated the enemy carrier when I met it.

Anyhow, here is hope that with delay they still get their homework done. If they get their acts together and heal the thing, then all will be forgiven and forgotten sooner or later.

Until then, my money stays with me.

[Arclight]

Think you're missing the point somewhat: a single Walrus can be ordered clear across the island and/or over rough terrain with trees and rocks without any problems.

The problem comes from the "follow" behaviour. A single Walrus following another does decent but will fall behind over distance. Having 3 following a "leader" just plain doesn't work.

As far as combat goes, they are pretty crack shots. Stationary targets are nailed in 1 shot, moving targets are trickier but dealt with efficiently.

The game works just fine if you give units orders individually, or at least break them down to pairs.


Seriously, try the demo.

[CaptainHaplo]

Skybird - let me see if this analogy helps...

You have 2 people. One is told to follow the other. This works fine.

You have 4 people. Three are told to follow the 4th. The three do not communicate. They thus either:

A) bump into each other and surrounding obstacles as each tries to be following the first person to the exclusion of the others

or

B) They keep one eye on who they follow and one eye on the others - meaning they now don't have time to look for all the terrain hazards.

Take your pick.
Now - as humans we can decide a "follow order" - choosing amongst ourselves who is going to "follow first". Such decisions are outside the scope of the "follow" command - because the gamer didn't tell player 4 to follow player 3 who follows player 2 who follows player one. Insert Walrus as player - and there you go. The gamer simply told unit 2/3/4 to follow unit 1. And they all do so - either tangling up in themselves because they were NOT told HOW to do it - or getting caught in the terrain while trying to be "first".

The complaint's your referencing amost all have to do with the same problem. Not pathfinding truly, but following. As Arcligh said - order a walrus across an island, and it will get there. Sure, it may not take the most ideal route, it may have to adjust on the fly a little. Oops - a hazard it didn't expect, so it backs up and turns and goes around it. How THAT is bad pathfinding I don't know. Is the admittedly "limited" AI supposed to know and see all navigation hazards at unlimited range?

The idea of a manta on autopilot just flying into a mountain? Really? I have never - from Beta 1 to patched release - seen anything like what you are describing. Could combat AI for manta's be tweaked? Sure. But I haven't ever seen a manta - on either "side" - just kamikaze into anything.
Having spend the hours of testing and then playing, I am going to say that is either a gross exaggeration that your repeating, or a fluke.

Regarding "following" - there is not a good way to tell your units to individually follow another. If you could chain the follow 4->3->2->1 then I suspect a lot of the issues would be solved by those players that want the AI to basically play the game for them.

Actual pathfinding has a few bumps, mainly because the Walrus AI is unclear on what it can and cannot traverse reasonably. But the complaints regarding real pathfinding are rare. The "follow" pathfinding is the major gripe - and I don't see it as a real problem because its obvious WHY its a problem. Not a programming issue as much as laziness on the part of the player.

I respect you want to hold onto your money. So go try the free demo!
Free is good!

[Skybird]

Haplo,

I refer not to my experience, but to the experience as reported by many people. The problem is not just 3 walrusses following the lead vehicle (they are a problem obviously, too, but not the only one), but the problem is that even the lead vehcile or a singular vehicle all to often messes up, gets stuck, circles around obstacles, rocks back and forth between two points, cannot catch up the next pathpoint it set because it set it too close, not recognizing bridges, or falling off from them, trying to climb too steep hills, getting bogged down by a single stair.

I linked one video. There is no second walrus in sight anywhere.

I cannot help it but many people report many more issues than just the little minor things that you are willing to admit. And many people said they have shelved the game because they consider the pathfinding issues a back-breaker. It also gets listed in quite some formal reviews around the web.

Nicetalking it does not help to motivate the studio to get the work done that it avoided so far. What they claim to do now should have been done already - several months before release, after the issues were reported after the first Beta.

Anyhow, nice to hear that you like it. I think the standards people expect to be met regarding quality in some regards, maybe just are different. I judge it by common, not too exotic standards of the market. And I think there are many titles out there that show how to do it better, and that it can be done better, and that a competitor aiming at the wallets of customers must do better indeed.

We have started to move in circles here now. So I pause until the feedback for the patched AI starts to come in. As long as that patch does not become a reality, I will not buy it anyway.

[Arclight]

Grabbed it. Love it. AI can use work but is perfectly servicable. Internet people are a bunch of whiny crybabies. News at 11.

[Skybird]

After a two weeks break from monitoring their forums, I went back in and looked and found this.

And this is rich. Really rich. They had this thread in their forum, just two posts long:

http://forums.bistudio.com/showthrea...dn-t-buy-again

Says the paying customer who correctly feels ripped of his money, is critical but remains cool but polite:

This game has crashed maybe 10 times on me today - every time losing ground and having to start back 20 minutes earlier or 10 minutes earlier.

Pathfinding is just...

Seriously BI, I didn't buy into a BETA.

This game was supposed to come finished. Sooner or later the gaming industry has to wake up an realise it's part of the rest of the commercial world - it's a lawsuit waiting to happen.

Won't be buying another half-finished BI game.

What a ripoff.

Replies the forum mod who represents BI's interests:

Thanks for registering just to rant. Do some research before buying a product and you would have known about specific issues. And you would also have known they are working on a patch to fix the most important issues thus saving you from registering and posting the (seemingly) useless rant.

And locks it.

A customer complains about having lost his money over what many see as a seriously broken product. The mod tells him it is his own mistake if he bought it in good belief and trusting the producer's self-advertising, and calls it a useless rant, then kicks him. Or do you read anywhere in the BI-adverts how lousy the state is they deliver the game in?

That's what I call good customer service and valuing the paying customer whom you expect to pay your income. First messing up and delivering a broken product. Then taking customer'S money nevertheless, and then holding the customer guilty for having been buying this unfinshed - thing...

I do not read on BI's webpage advertising that they deliver their unfinished thing it in a broken state and as a Beta. And when reading about the many, many, many things that indeed are reported and described as malfunctioning or broken, "broken" seems to be the only attribute giving a correct description of it.

That they plan to patch it, is no excuse. Things in such a lousy conditions should not be released int he first.

I have had the demo meanwhile. Painfully long 1.5 GB download. Then checking it. Terrible. It is an impudence to release stuff in such a condition. And maybe not clever to demonstrate it even in a free demo. Must be a decade or so since last time I saw something like this.

P.S. Anmd what did the mod mean when saying they plan to fix the most important issues - the rest of the mess stays uncleaned, or what?

[Drewcifer]

Its not a very good game that is a fact. I do regret the purchase but I bought into the beta so atleast I didn't pay full price. If I had a time machine I would go back and NOT purchase.