Interview with Ivan Buchta, your questions answered!

[radarhead 0]

At Gamescom we (The GameOverCast) had the honor of interviewing Creative director Ivan Buchta about Arma 3. We asked for your questions, and hopefully you'll enjoy the results.

I personally really enjoyed doing this interview, he was a great very personable guy.

Subjects covered:

-Ragdoll physics

-Ai improvements (cover usage)

-Aerodynamic upgrade

-Underground structures

-Mod support

-Six launcher equivalent built in to Arma 3?

-More realistic damage model for vehicles

-The campaign for Arma 3, will it be as open and hard to approach as Arma 2?

-Bi pod use

-Ivan's personal favorite improvement in Arma 3 over Arma 2

-Improvements and changes to the mission editor

-When is it bloody coming out!!!!!

And mucho mucho more.

You can grab it here: Episode 107 or on ITUNES 1 Hour and 14 minutes in.

Enjoy and maybe check out the rest of the show, its good honest :)!

We also will be giving away some Carrier command and Arma 3 swag in Episode 108 so stay tuned :).

[froggyluv 2130]

- Boooo to report any AI cover issues to bug tracker :mad: How about existing cover behaviour such as better deciding factors of when to stay or abandon cover? How about using non-map cover ie.. vehicles as cover? Also, as cool as underwater life looks -it seems they'd rather talk about the challenges of getting AI to function in this new environment over the already existing one.

+ Yeaaah to characters showing being hit rather then just red mist thanks to PhysX 3 :) I've wanted this for a long time. Hopefully physics system also handle falls (with stun effect) that deal appropriate damage/reaction from soldier.

+ Of course the HDR looks great as well.

Thanks for this interview all!

Edited August 26, 2012 by froggyluv

[-Coulum- 35]

Good interview although it is kind of disappointing there is no concrete plans for weapon resting. Oh well, ace will come to the rescue hopefully.

Happy about physX actually making hit reactions on soldiers rather than the current robots that don't even flinch when hit.

Thanks alot guys.

[Fireball 16]

- Boooo to report any AI cover issues to bug tracker :mad: How about existing cover behaviour such as better deciding factors of when to stay or abandon cover? How about using non-map cover ie.. vehicles as cover? Also, as cool as underwater life looks -it seems they'd rather talk about the challenges of getting AI to function in this new environment over the already existing one.

Here are a selection of related tickets to above point - most being there for a long time already:

https://dev-heaven.net/issues/25544

https://dev-heaven.net/issues/25543

https://dev-heaven.net/issues/25060

https://dev-heaven.net/issues/24113

Another thing that makes me wonder (pisses me off a bit) is that the weapon resting is low priority on the wish list, but in the sentence about weather in the question just before, he mentions that they want to keep it to the "technical military simulation" aspect - which definitely would need weapon resting. I mean, looking at comparatively trivial shooters like Battlefield 3, which show how it's done right, you should think BI should have gotten the hint by now.

[maturin 12]

Anyone who asks about bits falling off vehicles should have a tire full of burning gasoline dropped around their neck.

Armor penetration, HEAT, KE, ERA, spall goddamnit!

Cool news about a SixUpdater analog.

[chortles 263]

Anyone who asks about bits falling off vehicles should have a tire full of burning gasoline dropped around their neck.

Inciting violence against ARMA players? Oh boy :rolleyes:

Fireball, re: the "low priority" of weapon resting but "wanting to keep the technical military simulation aspect"... I would not be surprised if the BI/ARMA 3 dev team vision of what "technical military simulation" means does not view weapon resting as definitely needed.

Then again, I personally suspect that the unspoken message is "we're not going to add features until the game is so stable that a casual wouldn't complain".

[arigram 0]

Then again, I personally suspect that the unspoken message is "we're not going to add features until the game is so stable that a casual wouldn't complain".

You are most probably right.

I am certain that the focus of Arma 3 right now is the new core engine features, such as the PhysX integration, animation system, new lighting and general stabillity and many minor features will be dealt upon when those are in good shape and there is enough time and will to invest.

[.kju 3239]

Thanks for the interview and posting here radarhead :bounce3:

[krazikilla 5]

So basicly nearly all of the points got answered with a: Nope sorry, all we do is finishing the new visuals (graphics, lightning, ragdoll, physix, chemlights!!)

We would like to have new featuers in, but graphics/physix are first.

Means:

either they put in some features from whislit at the end of the development, and delay the game for this cause 1-2 month.

OR, they dont care bout most wishes and release it after polishing physics and graphics with nearly the same gameplay as arma2 :)

Edited August 27, 2012 by KrAziKilla

[MavericK96 0]

Then again, I personally suspect that the unspoken message is "we're not going to add features until the game is so stable that a casual wouldn't complain".

Which is a noble and welcome goal, to be sure. But not if that means removing/neglecting a bunch of features that dedicated ArmA players have been waiting for.

[chortles 263]

Does it count as neglect if they have differing priorities than dedicated ArmA players? The approach already reminds me of what InstaGoat says he heard from BI at Gamescom when asking about the devs' approach to AI.

[old_painless 182]

I think it was RiE who said that they did want to fix the AI, but the code was not well structured and built on many contributions from developers who were not with BI anymore. So I think they fix and improve what they are able to do well, and not open up the can of worms that AI is. Too bad, but that seems like what their mindset is.

So worst case scenario is that in Arma3 we will be shooting at AI teams across beautiful landscapes and see the first guy go down in ragdoll glory while his other three AI buddies just stand there next to his trembling body. Let's hope that they will be able to manage at least some AI improvements

-OP

[Tonci87 163]

I think it was RiE who said that they did want to fix the AI, but the code was not well structured and built on many contributions from developers who were not with BI anymore. So I think they fix and improve what they are able to do well, and not open up the can of worms that AI is. Too bad, but that seems like what their mindset is.

So worst case scenario is that in Arma3 we will be shooting at AI teams across beautiful landscapes and see the first guy go down in ragdoll glory while his other three AI buddies just stand there next to his trembling body. Let's hope that they will be able to manage at least some AI improvements

-OP

The Problem seems to be much more complex. It seems that AI was build and improved without proper documentation. The Devs seems to have problems understanding the AI code. Their goal for Arma 3 seems to be to complete the documentation and to rework the code. That will allow them to go deep into the AI routines and to make very noticable improvements. Unfortunately this process will take some time and most likely won´t be finished until A3

[chortles 263]

Here's the only post I found by RiE on AI:

I think the point is not that more 'deep rooted' improvements to the AI aren't warranted/ desired; rather, it's that to be confident about which behaviours need to be changed/ corrected, one must be sure that the configuration is balanced appropriately. It's also worth bearing in mind that more high level observations about AI decisions are not simple to translate into technical requirements.

Naturally, I wouldn't dare to suggest that specific repros aren't available (The CIT and Beta Patch forums have been and continue to prove to be invaluable resources for our most skilled programmers: the ones with the experience to make core changes happen), but, again, we need clean repros and it can be a challenge to achieve that at this stage of production, where a big number of factors interplay (i.e. significant environment, model, config, animation changes on a frequent basis) on top of the significant resource demands of implementing new features.

Looking forward, we know that AI refinement will be necessary. The above isn't intended to excuse not achieving it; rather, it frames the problem and doing that helps us plan an appropriate schedule for getting it done. :)

Here is Old Bear's report on Gamescom:

On the question of "AI cheating with grass," Ivan told me that at the moment nothing had changed, they had tried several solutions without success, until now the solutions penalized too much the IA operations, and at the present time, rather than setting up a complex set of AI tweaks with multiple cursors in lieu of single "skill" cursor, BIS is moving towards a system limiting detection modes/perception by AI.

Regarding the 3D editor, there is no progress there because BIS does not have the means for it. All scripters being mobilized for the implementation of PhysX 3, for the animations, the AI and the GUI ... and especially for the integration of all these elements.

In fact, if I understood correctly, BIS made lots of tests that work more or less well. Ivan told me that for example in VBS, they found a way to include objects underground to tunnels, passages ... and it works in VBS, but in Arma3 environment , AI perceives the object hidden in the ground as a barrier and tends to avoid it... which is not the goal .... Furthermore, the IA refuse to engage in tunnels created. So for the moment, no underground objects in Arma3.

Finally, here and here are InstaGoat on the AI as of the Gamescom build:

Sound is tip-top, aside from a few samples that I like because of personal taste (handgrenade explosions for example). I asked also wether I could have a look at the editor, but that wasn´t possible at that point unfortunately. However, I learned that they are working on improving the existing editor to make it work better, similar to the way they expanded the Editor in TOH, except specifically geared towards Arma (obvsly). The 3D editor will stay for sure, it seems, so we can stay hopeful that once time and resources become available, it will start to be developed further too.

What I learned about the AI was pretty surprising. The gist of it is that previously, the AI was being expanded instead of being properly configged, while now they are specifically working on making the configs work out so the AI stop making baby mistakes. (Snipers getting stuck with their handguns out, or soldiers refusing to throw grenades, for example. The AI in A3, as far as I could see, now use grenades very often when they are trying to get at an opponent they can´t directly see.)

Well again, basically what I initially said. I was told that one problem with developing the AI so far has been that new features had been added, while old ones basically were carried over. As old developers went and new ones came, documentation was neglected, so at times, the people currently working on the AI themselves don´t know what part does what. The result of that is the AI in Arma 2 is as clunky as it is sometimes: it´s not because it is bad, it´s just that it is made of lots of parts that were left unadjusted. So, what they are doing with Arma 3 is apparently one big spring-clean trough the AI framework, building and confirming documentation, constructing up-to-date tools to adjust and test the AI, and iron out the bugs.

That said, what I saw from the AI was okay, solid work as far as Arma goes. They react not faster, but smoother and more consistently. I didn´t see -much- though, and I couldn´t do any controlled tests like I did with my AI boxes I have on youtube, so the verdict is still out on this one.

Oh, and Gaia (apparently InstaGoat's main source at Gamescom) also told another interviewer in another video that the AI will not autonomously do unscripted minelaying.

[maddogx 13]

Oh, and Gaia (apparently InstaGoat's main source at Gamescom) also told another interviewer in another video that the AI will not autonomously do unscripted minelaying.

And thank fuck for that. When and where to lay mines is extremely dependent on the situation / context - something only the mission designer can provide.

[chortles 263]

I think Gaia said that the AI was more likely to simply break the mission -- thus ruining the fun for the players (hint hint) -- if the AI went autonomously minelaying on its own.

But yeah, basically no commentary in the previous post of mine, just wanted to directly quote so that you guys could have at it, unless you'd rather just do so in the AI Improvement thread, seeing as the gist of the response to the thread topic (Lord Ivan's interview) was "they're not catering to MY priorities".

[nik21 1]

Can anyone say something about the damage model of vehicles? Don't wanna download the whole interview with my slow connection. :)

[maturin 12]

A 'lay mines' waypoint would be nice, though.

[Undeceived 392]

A 'lay mines' waypoint would be nice, though.

Yeah, good idea.

And also, I suppose (resp. HOPE) that as commander we can order the AI do lay mines, just like we can order them to lay a satchel.

[BasileyOne 10]

can't wait to test Arma3 ;=)

cool news then.

generally more concerned about performance issues we're had in Arma2 anytime here[from beginning].

offloading GFX stuff/workload to GPU would help considerably then.

obsessions with supporting DirectX9[or DirectX10, which is actually as much deprecated/obsolete as DirectX9 two years ago], cause serious issues there, sadly.

another key thing would be moving from thread-based parallelism to message-based.

look at results! Yaws and Mochi beating Apache about 500x times[unsure how it against NgNx] on same hardware, literally causing kernel panic on old kernels/systems ;=)

SimpleDB, RabbitMQ and other suff, most people use everyday without knowledge about underlying stuff.

basically its not need to learn Erlang for quick start, few libraries for porting intrinsic/hot-pieces of appz could b temporal shortcuts b 4e learning.

[chortles 263]

"A 'lay mines' waypoint" and ability to order AI in your squad are possible but not confirmed, Gaia only stated that the AI won't be minelaying without some sort of external trigger.

[roni 11]

Can anyone say something about the damage model of vehicles? Don't wanna download the whole interview with my slow connection. :)

+1 !

Everyone talk about upgrading AI(i think it gets really better after few latest patches) and graphic effects, but what about the most important part of battles? Why it's so neglected, why noone want to answer about this important aspect, since OFP almost nothing changed. Why i think it is so important? Because whole result of battle depends on it. So what that I have M1A2 abrams, when it always destroyed after f.e. 3 hits from T-72?? And it even doesn't matter what part of my tank it hits, what is the worst thing..

I mean, what the fuck devs, we have 2012, games from '90 have better damage systems, you must do something with that!

Sorry for little O/T.

[chortles 263]

+1 !

Everyone talk about upgrading AI(i think it gets really better after few latest patches) and graphic effects, but what about the most important part of battles? Why it's so neglected, why noone want to answer about this important aspect, since OFP almost nothing changed. Why i think it is so important? Because whole result of battle depends on it. So what that I have M1A2 abrams, when it always destroyed after f.e. 3 hits from T-72?? And it even doesn't matter what part of my tank it hits, what is the worst thing..

I mean, what the fuck devs, we have 2012, games from '90 have better damage systems, you must do something with that!

Sorry for little O/T.

Here's a probable answer: they may well have their priorities (did I seriously have to say this again in the same thread?) and may very well have different ideas from you of what is an "important aspect" behind the concept of ARMA 3 (that is, "what justifies a new game instead of simply making more ARMA 2 DLC").

Not to say that yours isn't a valid complaint, but we have yet to see what it'll look like by the time we get to play it, and they may have different things that they want to deal with first.

I vaguely recall Jay Crowe mission-killing a wheeled vehicle with 6.5 mm rounds in the "infantry combat" GameStar interview (7/7), can anyone confirm or deny?

[nodunit 397]

There was also early footage of a wheeled vehicle being shot in the engine by an anti material rifle, causing it to come to a complete stop. Light armor has had penetration since at least OA, if you test with the HMMWV's you'll find that the up armored cannot be pierced by anything other than a heavy machine gun, whereas lighter ones such as the UAV and normal humvee's can be shot through by rifles, case and point I once shot the rear of a UAV HMMWV, the bullet went through the back and the trajectory sent it through the drivers head.

Armored vehicles on the other hand have no penetrating values, or more importantly bounce values, angles of armor don't seem to matter regardless of how effective the penetrating head of the projectile is or angle of the armor, making angled assaults only matter in crew awareness and that one extra hit vs bouncing off the strongest armor/most angled. I hope the introduction of PhysX will make such things possible one way or another, along with traction and grip from the caterpillar tracks..good news on that though is that we've seen drifting and being able to drive over obstacles in some of the more recent A3 videos so I have high hopes in this regard.

EDIT: Aw dangit now I look like an illiterate.

Edited September 5, 2012 by NodUnit

[Iroquois Pliskin 0]

There was also early footage of a wheeled vehicle being shot in the engine by an anti material rifle, causing it to come to a complete stop. Light armor has had penetration since at least OA, if you test with the HMMWV's you'll find that the up armored cannot be piercing by anything other than a heavy machine gun, whereas lighter ones such as the UAV and normal humvee's can be shot through by rifles, case and point I once shot the rear of a UAV HMMWV, the bullet went through the back and the trajectory sent it through the drivers head.

v7Q_1Pd9ME4

Engine block shot at 9:50, this is from last year's Gamescom. Great progress in retrospect: the above demo looks like ArmA II on an island like Lingor - not much of an improvement back then, compared to this year's E3 & GC.

They seem to be upgrading the damage model from game to game: does anyone remember what happened when you shot the wheels in Armed Assault 1? LOOOL.

Armored vehicles on the other hand have no penetrating values, or more importantly bounce values, angles of armor don't seem to matter regardless of how effective the penetrating head of the projectile is or angle of the armor, making angled assaults only matter in crew awareness and that one extra hit vs bouncing off the strongest armor/most angled. I hope the introduction of PhysX will make such things possible one way or another, along with traction and grip from the caterpillar tracks..good news on that though is that we've seen drifting and being able to drive over obstacles in some of the more recent A3 videos so I have high hopes in this regard.

Adding separate ammo and crew compartments would be a major upgrade. I'd like to see scenarios, where the crew dies without the tank being completely destroyed, or a tank becoming totally disabled without it going up in flames, unless the ammo compartment is hit.