New MMOFPS from NetDevil:

[Venkman]

Article. Announced at CES. Wish my travel schedule would have allowed me to go this year. With every show redefining themselves once E3 changed, I think there'll be some goodies at EVERY show in 2007.

Anyway, MMOFPS, Unreal Engine 3 using the AGEIA PhysX card thing I'm sure everyone here bought already, right? RIGHT?!

Sounds like:

• Ranks to unlock new abilities.
• Destructible/deformable terrain.
• Actual Collision Detection. Really. They mean it

I like it. They may not be the best at creating compelling retention-based MMORPG, but the driving portion of AA was fun as heck. Maybe this can be that plus, like, a reason to stay after the free period.

The year is 2029 and the Global Corporate Wars have begun! Warmonger is set against an apocalyptic setting, when two of the largest military spenders, PolyChem Oil and General Energy, clashed in a dispute over a large cache of Iranian oil fields. When the conflict escalated, an all-out war took shape on US soil. The game is staged for players to take down an entire city; one block at a time. As maps are won, a larger tactical influence is then triggered in the next, or surrounding, map(s). Each map instance will play differently as the sheer destruction of map elements forces players to adapt and find new ways to win or defend their objectives. Every round leads you closer to dominating the web of maps that make up the entire city and players have the ability to raise their ranking in a class, gaining extra abilities for high performance. Operation: Downtown Destruction is part of a longer series involving distinct periods from the episodic Warmonger story.

Unique to Warmonger is a destruction system that allows dynamic gameplay to emerge from the results of combat, where every aspect of the environment can be completely leveled. Destruction is done procedurally, rather than pre-canned animations that are found in most games today. As a result, explosives, rockets, indirect fire, and vehicle combat can literally blow away walls, drop ceilings and open up new pathways for enhanced gameflow. Collision detection has also been carefully considered during development, and the effect of destruction within the environment can be used as a weapon. A sniper can blow the stairs behind him to block access, but a rocket to the floor beneath him will drop him down, causing possible death.

[Trippy]

Who TF is giving these guys money to make another game?

[Venkman]

People who only played the driving portion of AA and the flying portion of JG :)

[Yoru]

Wake me when the beta starts. Until then, it's all just pretty words.

[WayAbvPar]

Who TF is giving these guys money to make another game?

YA RLY. Auto Assault wasn't enough of a debacle?

[Simond]

(insert War/Whoremonger joke here).

[Trouble]

I like the premise a lot.

[Chenghiz]

Kotaku's post regarding this leads me to believe it'll require one of those physics processors. If so, that stinks.

[schild]

This premise is too good for NetDevil.

Edit: Also, Crecente & Co. couldn't tell the difference between their ass and a hole in the ground. It won't require a physics processor. Guaranteed. I'd bet on it.

[Ratadm]

They screwed up post apocalyptic cars lets see how they can screw up an fps.

[slog]

Netdevil put all the hit detection on the client for Jumpgate.  Also, if you played on a really laggy connection people couldn't track you so you could fly up and hit someone over and over without being detected.  2 seconds later you start taking damage with no idea knowing who is hitting you.

Also, the original server for Jumpgate ran Windows 98.  (source on that = GreatBob).

[sinij]

This premise is too good for NetDevil.

Premise is too good for anyone at this time.

There is a reason why real-time deformable terrain is not done in mmorpgs - and that reason is bandwidth requirements between server and client and computational power required to render terrain real-time on client's side. I'd say this will not be feasible for another 10 years.

[geldonyetich]

They say the third time's a charm.

[Endie]

Yay, at last I get to live in the world of General Energy and tell my own story alongside literally several other players!  I hate those PolyChem Oil bastards.

Actually, it sounds kinda fun.  But that's probably because I'm imagining some dreamy, teamworking, non-schizo version of MMOFPS reality where it's not either a zerg rush or a taxi to victory scenario.

[Ironwood]

Also, the original server for Jumpgate ran Windows 98.  (source on that = GreatBob).

That's gotta be a windup.

[Falconeer]

Still think Auto Assault could be mildly entertaining if free. I mean, Dungeon Runners fun (that kind of "mildly").

[Endie]

Don't derail this thread.  I want to hear more about taking cities block by block in Stalingrad Online.

[eldaec]

Two is the wrong number of realms.

And I'm dubious that terrain will be as deformable as they imply in the end.

Other than that - hooray for content free releases.

[Modern Angel]

They screwed up post apocalyptic cars lets see how they can screw up an fps.

Exactly. They took the greatest idea for a MMOG ever, one that couldn't possibly fail (CAR WARS WITH CYBORGS), and made it fail. Fool me once...

[Stephen Zepp]

This premise is too good for NetDevil.

Premise is too good for anyone at this time.

There is a reason why real-time deformable terrain is not done in mmorpgs - and that reason is bandwidth requirements between server and client and computational power required to render terrain real-time on client's side. I'd say this will not be feasible for another 10 years.

Sorry, but that's simply not true ;)

--if you are networking the entire terrain, sure networking performance is an issue. Instead, you network deformations, and have the clients apply those deformations to the base terrain(s) downloaded, and/or periodically synchronized base terrains updated when you log in.

--most terrain systems already render, and in fact generate on the fly the vertices to physically render. Even leading edge tech has a series of mechanisms where actual rendering of the geometry isn't your limiting factor.

I wrote a networked terrain deformation object for Torque's heightmap based terrain engine (TGE back in the day) before I was even an employee. It's been used in two games to date, and the theory has been applied to a couple more in development now.

EDIT: I went back and read the linked article, and from what they are describing they mean destructible environments not just terrains.

This is more difficult, especially once they realize just how difficult it is to keep clients synched when using hardware physics, but it's doable given an outstanding design and careful limitation of what is gameplay affecting and what isn't.

Sounds like they want everything gameplay affecting, so this will be interesting to see!

[HaemishM]

What makes anyone, including the functional rich retards who gave them the money to develop this, think that NetDevil can make a good fucking Pong MMOG, much less an MMOFPS? They made Webster's add another line in the dictionary for the definition of inept.

[Krakrok]

AA already has destroyable environments. They're just using their existing engine. The physics card just gives them more random debris. It's also an option when you customize a Dell.

If they have 10k subs on AutoAssault they're pulling in $120k a month. The real question is how much did it cost to begin with. If they are using their existing engine to bootstrap the new game they could have it done in much less time (or not) on a smaller budget.

[Trippy]

AA already has destroyable environments. They're just using their existing engine. The physics card just gives them more random debris. It's also an option when you customize a Dell.

If they have 10k subs on AutoAssault they're pulling in $120k a month. The real question is how much did it cost to begin with. If they are using their existing engine to bootstrap the new game they could have it done in much less time (or not) on a smaller budget.

They (as in NetDevil) aren't pulling in that much since NCsoft funded them and is presumably taking in the bulk of that money.

[Krakrok]

Grossing ~$120k. Who knows how much it cost initially ($2-5 million?) or if NCSoft is getting a %. AA uses Havok 2 and already supports the physics card so really nothing new here. 'Warmonger' seems to be some kind of quest in AA that they are carrying over to it's own game. The physics card company could be paying for the new game but I don't see any press releases to that effect.

[UnSub]

It kinda sounds like a sub-game suggested for AA that is being spun off into a whole new game.

Also, to make the terrain deformation easier to deal with, I suppose they could instance or pseudo-instance the blocks - after all, once you've won a block, you probably won't be going back into it so it can remain static and closed.

Could be interesting.

[sinij]

Sorry, but that's simply not true ;)

I'm curious how would one go about introducing fully dynamic terrain environment, something that you can't just script in a form of 'big hole"/"small hole"/"trashed building" ?

[Viin]

Still do'able, but it depends on how cheesy they want to get. It's very possible there are only certain things that are destructible.. in which case they just have to send a 'this item now destroyed' packet to everyone rather than actually doing anything is destructible. This is how Auto Assault works anyways.

[Krakrok]

Right. You just cheat and say blow this item up. Everyone sees a different explosion but doesn't know it.

Sacrifice had deformable terrain in a networked environment (and so does Second Life?). You could cast a spell in Sacrifice and a volcano would grow up out of the ground.

[Stephen Zepp]

Right. You just cheat and say blow this item up. Everyone sees a different explosion but doesn't know it.

Sacrifice had deformable terrain in a networked environment (and so does Second Life?). You could cast a spell in Sacrifice and a volcano would grow up out of the ground.

All the difference in the world lies in if you want the effects of the environment destruction to have game play value.

Just about all players, and a vast majority of developers, go down this path of thought:

1) I have a phsyics solution/card!
2) I can blow things up and have them create new debris objects, that can then be used to do nifty stuff!
3) Since I have the same physics card on all clients, they can all see the same thing at the same time!
4) Since all clients (and the server) all see the same thing at the same time, we can do all sorts of nifty stuff, like have explosion create flying debris that knocks into players, and lands on the ground and forms new obstacles, and, and and!

The problem is that many modern physics libs (and right now I will put a big fat disclaimer up: I do not understand the details of how these cards solve their physics systems, just what I've been told by those that do know) do "total solutions" all at once--in other words, they don't solve a subset of the physics simulation a piece at a time, but generate a solution for the entire set of data at once (something to do with esoteric math from what I understand).

What this means in a practical sense is that if any of the data points within the simulation are even slightly different, the total solutions could wind up being drastically different. You can't simply say "do this type of explosion with these parameters" and expect your clients to see the same result unless your server and your clients are completely synchronized--which implies zero latency and nearly infinite bandwidth. In other words, it's not gonna happen.

The only real workable solution given today's capabilities is to have the server handle all gameplay affecting physics, and then ghost (network) down the resultant changes to the clients (normal game networking)...but that means that the client isn't going to be able to very accurately predict what is going on while the player is doing his reaction to the event, and therefore will see a short period of what the client predicts is going to be the result of the explosion, and then a correction period where the server tells the clients what actually happened. Due to the problem described above, the physics card cannot help with this at all on the client, and if you try, it makes the corrections worse, not better.

Unfortunately, that leaves us right back where we were--either limited due to networking bandwidth regarding exactly what can be game affecting, or what could be gross inaccuracies in player viewpoint until the network updates settle down. And that's just not what people (players, designers, developers) want.

Currently (and GG has put a lot of R&D into this), we're basically stuck with having client side physics cards able to do some amazing "eye candy" affects, and for single player games some really cool game play affects as well, but networked game play events are still limited in their total use of client side physics libraries, especially "total solution" libs/hardware.

I'm not saying it'll never be solved, and I'm sure that smoke and mirrors combined with some really accurate prediction algorithms can make it look pretty good, but today, networking and physics is like oil and water...

[Typhon]

3) Since I have the same physics card on all clients, they can all see the same thing at [some point]

Not ideal, but it seems like a decently acceptable next step.  Specifically what I'm talking about is; when a destruction event occurs, hold off the animation until the responsible computer (I assume server, but I don't know if putting a physics card in a server makes sense) has computed and communicated the final solution.

[Stephen Zepp]

3) Since I have the same physics card on all clients, they can all see the same thing at [some point]

Not ideal, but it seems like a decently acceptable next step.  Specifically what I'm talking about is; when a destruction event occurs, hold off the animation until the responsible computer (I assume server, but I don't know if putting a physics card in a server makes sense) has computed and communicated the final solution.

But they never, ever will.

By definition, in a server authoritative setup (which you need for anti-cheat), the client is always lagging behind the server in world state, so to stay reactive to user input, it must predict where it thinks the server's updates are going to place objects, and start them along the way. Then, when the authoritative updates show up, it gently corrects them in accordance with the information in the update.

This basically means that with a player making constant inputs, the client side simulation will never be 100% sychronized with the server side simulation--it's a constant tug of war between where the client thinks everything is and where the server tells the client where everything was a bit ago in the past...

[sinij]

Can you please explain why having exactly same explosion is important? You blow something up, you get a lot of eye candy and then server updates your client (oh btw you got killed by flying anvil) with results. Big explosions take some time to 'play out', you might have enough time to update clients with significant outcome. That and you can tell client to show all or most possible outcomes that will end up as false positives ('eye candy' effects) when server updates with real data.

[Viin]

I don't think it matters - as long as the end result is the same for everyone.

The problem is that that's not a true 'deformable environment' if the destructible stuff is predefined.

[Lantyssa]

Can you please explain why having exactly same explosion is important? You blow something up, you get a lot of eye candy and then server updates your client (oh btw you got killed by flying anvil) with results. Big explosions take some time to 'play out', you might have enough time to update clients with significant outcome. That and you can tell client to show all or most possible outcomes that will end up as false positives ('eye candy' effects) when server updates with real data.

The problem is that the side of the building which just cleaved off is now part of the terrain.  If it falls straight down on one client, but tumbles then lands twenty feet away on another client, synch them back is going to look very strange for someone.  If the server decides something else, it is going to look odd for everyone.  Instead of people sliding across the landscape, the landscape itself is jumping all over the place.

One way they could hide some of the predicition and synching is by producing a lot of smoke, eye-candy debris, dust, and fireworks.  Really though, the obfuscation would have to be overdone and a significant portion of using the physics card would be lost.  (At least the explosions would be pretty.)

[sinij]

Hey I will take smoke and mirrors, or even scripted destruction, over what we have now - nearly nothing.