Is twitch based combat feasible for MMO's yet? - FSOD

FSOD = For Sake of Discussion.
*I am not developing and MMO, nor do I have any immediate plans to. This discussion is for the sake of discussing.

WoW will always hold a place in my heart for its over-all story and shiny polish. GW2 and WildStar both attempted at more skill based combat systems, but they still fall prey to the clunkiness of targeting and spell casting. TERA made an effort, as have Mortal Online and DarkFall.

Unfortunately, at least to me, none of the systems come close to the feeling of a good ol' FPS match or solid melee feels of a game like Chivalry. MMO's feel slow and require a great deal of lag-play and/or button mashing to get the desired effect. In the end MMO's always fall back on gear levels and stats, with only the elite players getting to experience a half-arsed skill based system (still have to deal with lag even if you're l33t).

Are networking systems/infrastructures anywhere near having the capabilities required to give a truly massive MMO a quick twitch feeling like that of dedicated server pvp games?

The main idea in my head that spawns this question is: An MMO where no levels or numbers effect PvP, only player skill in a real time system. Levels could be left to crafting. (But even then I would want to avoid numbers all together if possible, outside of base things like HP, creating progression based on discovery, player driven economics, player driven politics, etc... Think WurmOnline sandbox, Mount and Blade combat, etc.) Don't worry, this is just an idea in my head that I wanted to express as an example of how well functioning twitch combat could change the MMO-sphere away from WoW clones.

 

Dungeons & Dragons Online encourages a bit of movement to avoid incoming attacks. Ranged attacks especially but you can somewhat avoid incoming melee attacks. It isn't always feasible to do so though because the servers tend to lag at times. That's mostly the fault of Turbine having to work with an old code base (DDO is closing in on nine years).

 

Isn't that what Planetside is... an FPS MMO? It's got a sequel too. Also Defiance, which is a third person shooter?

 

Suppose I should have been more specific, or followed my 'MMO' with 'RPG'. I'm talking outside of the FPS genre.

Planetside2 is definitely a move and shaker in the MMO world.

 

How does changing the genre matter? You asked if it were feasible to achieve and clearly it is.

 

Its not clear yet. Even PS2 has it's issues, and an FPS doesn't necesarily prove good melee/magic/etc. combat that an MMORPG might have. Plus balancing the crafting in a game where combat is based on real time.
This is meant to be a discussion about possibilities not blanket statements.

Though PS2 is definitely makes a strong argument for more twitch based MMO's.

 

What you described, namely trying to find good gameplay mechanics and balancing crafting, has nothing to do with server-side capabilities. You asked if it were feasible and PS2 clearly shows it is.

 

Really? Because you also said this in the very first post...

"Unfortunately, at least to me, none of the systems come close to the feeling of a good ol' FPS match "

 

Perhaps it would help if you provide some examples of they type of gameplay you're thinking about not found in MMOs as you see it...

My first thought, when you used the word "twitch" is the near instantaneous response time you get with consoles et al. There are 2 key elements to that discussion in an MMO 1) the standard gameplay and 2) the technical implications. As is often the case in "pushing the envelope" the only way to achieve the effect is to push the later as far as possible, then use the former to compensate for whatever is left. Oh, and once it's done... you'll be hard pressed to change it (ever). The WoW team can only "innovate" so much before large numbers of their dedicated users (read paying customers) will cry foul (even if the change is an improvement... that's a psychology thing I won't dive into now).

Technically to create a MMO you'll need to synchronize the various players presentation at least every 4 frames. Transmission, serverside processing and broadcast reception will take the better part of 2 on the fastest networks on earth. Then your player system will need to process what's changed... on a lot of computers and systems, you're already outside your 4 frame window. That's why you need to use predictive techniques, which are often wrong unless you manage player input to prevent sudden "unexpected" behavior. Enter game design to save the day

Twitch gameplay, in my experience, is near instantaneous "arcade" like experiences. A lot of games, including FPS and others, sacrifice true twitch response to allow for complex processing (like advanced AI) and high end graphics to perform smoothly (particularly on less than the bleeding edge hardware).

If you're looking for a player "dodge" effect, the best gameplay mechanism that jumps to mind is time dilation. The act of shooting causes visual "slow down", while the defensive player can move at normal speed.

Hope that helps,

Cheers,

Galen

 

@Mikenseer - Like @Ryiah is trying to tell you, the genre doesn't matter when discussing twitch-gaming in a MMO environment. The main bottleneck is technology in this situation, Planetside 2 and DC Universie is the best commercial examples that I can think of at the moment. Maybe Landmarks and EQ Next is even more interesting because they seem to head in the twitch direction as well but they also have to communicate voxel data between servers and clients.

There are some research projects that boast about reaching even better results but I don't have first hand experience with them and therefor take their results with a bag of salt. The reason for the skepticism is that I am wondering who it is that finance these... MMO grade servers aren't cheap to run...

That being said... the main problem with twitch-based game mechanics in MMO environments is that they are more expensive to calculate on the servers. The second problem is all the connections that have to be kept and managed between the server and clients. The third problem is the connection speed of all the clients... they need to regularly update the world state... not just players, npcs and dynamic assets... now the projectiles also have to be synchronized. It can be done but it increases the information that must bet transferred between the client and server... imagine a few hundred players going nuts, fighting each other... jikes

Sony has solved this, probably with ridiculously optimized packages and server/client code. But you can't get the level of precision that you see in some 32 player, fps games... that would just be like asking for the impossible at the moment but it will com in the future.

 

It's perfectly feasible in this day and age for people with decent connections, however the problem is it's quite expensive to run twitch gameplay servers due to vastly increased traffic, and the requirement to have a lot more servers locally to people.

Lets consider how WoW works: it collates all your decisions into regular packet cycles, and these pump out at an estimated rate of say 2-3 times a second. Twitch gameplay requires you shoot and the packet is sent almost instantly. While this doesn't seem like much more of a big deal, it's a big deal server side processing it and with bandwidth.

Determining collision resolution with authoritative server demands much more.

But I don't see any real reason why you wouldn't other than MMOs currently get the majority of their business from casual gamers. If it only catered for the hardcore, and it was popular, it would cost much more.

 

@hippocoder - I agree it's possible to come real close (at least from a player's perspective), but again I'd assert "twitch" gameplay in the console/arcade sense is a 1-4 frame response time to user input. For the types of games that typically use that style of gameplay, even at 4 frame delay between action/response being users may notice a response lag (more like a slight stutter I grant you, but still).

Basic networked games (and Unity's default) use 15 frames (4 times a second @ 60fps targets). Every 10 frames would put us at 6 times a second. Full round trip and display should be possible within 10 frames for good connections/equipment, but without tweaking the player's expectations (via plausible gameplay mechanics and story setup), i'd still maintain it'll never stand up to the scrutiny of dedicated twitch players.

 

MuchDifferent with their Unity Park Suite claim to have tech that allows 1000 people in a single FPS battle, but ... that's demo tech in a controlled environment.

 

That being said, I think GuildWars 2 (GW2) is more successful at a compromise between twitch and static game play.

There is a "lag" between mashing a button and the delivery of a specific action, but this is part of the Game Design; and more than just adjusting for servers determination, etc..

As part of the game, actions in GW2 take time and have associated animations that match that time. If one's character does an action:

Spoiler: View Image

... not only does the player need to wait for the animation to end:

Spoiler: View Image

... but also needs to wait for the time allocated in Game Balance for Damage Per Second, cost of dodging an attack, or whatever the designers think is necessary for the game to work.

Now, this *can* lead to button mashing... and perhaps there is some weakness in design where one ranged attack becomes the most optimal for too many situations, so the player finds themselves mashing the same button...

... but I think this is a difficult and somewhat inherent part of the tradeoff to making games.

Personally, I find the compromise acceptable.

In Day of Defeat (DoD - my fav. twitch game so far...), I ran out of breath when sprinting. In GW2, I can only dodge twice before I need to recover. In DoD, I have to wait for my gun to reload, or the bolt to be cycled. In GW2, I have to wait for my animations to finish playing or my cooldowns/talen bars to reset.

I see this as much part of the intentional GameDesign as anything to do to cover holes or flaws in the technology.

I don't see anything from stopping a designer from creating a story based game with deep narrative content, where the battles are based on quick short fast attacks.

This just may be more limiting to the game play if the design prohibits anything that takes time. Even in DoD, there was the need to plant mines or time consuming actions.

 

To be fair its not a twitch fps

 

The real solution to what your saying, and probably the "holy grail" of MMO's, is a server side MMO.

Everything is done on the server side.

EVERYTHING.

Then have a rendered video stream sent to the client and only take controller input back from them. This would eliminate all cheating and hacking, allow almost infinite world detail and infinite players on one "sever", and it would allow the game to be run on almost any system with a descent internet connection.

That or you could do a dynamic drop in and out type server like Destiny.

 

Journey had a very nice drop in system

 

IMO twitch combat is just not something most people who play mmo's want. Over the years I've played a lot of the massive pvp combat mmo's, and almost universally the people who play those don't really like twitch play. They like the team oriented combat, the higher level tactics and strategy, but not twitch.

Technically you could add a lot more twitch then you currently see. Most mmo's don't optimize heavily for space. That happens to be something I've worked on extensively and games only generally go as far as they need to, and for mmo's it's not really that far. Most do basic bit packing, but you would be surprised how many send floats instead of ints and then converting, and it's rare to see an mmo using stuff like delta's, even though it's a rather straight forward thing to do. GW2 has some of the best optimization I've seen to date.

I've done hard testing on this and with modern broadband connections you can have around 200 players or so in visual range and support a lot of twitch combat before you start to overload the client bandwidth, which in the end is the hardest problem here to solve. Clients run out of bandwidth way before servers have trouble keeping up, and by a huge margin.

While it varies most pvp centric mmo's use somewhere around a 60ms tick interval, so you are getting location updates around 16 times a second. And what that number could be is highly dependent on your space optimization. I'm running at 30ms ticks and I'm more efficient then most games at 60ms, but I heavily optimized for space.

 

And if Henry Ford asked people what they wanted they would have said "a faster horse".

 

I bet that before Game of Thrones most writers thought "killing off every likeable character just isn't something people want", too. And it isn't. But the series is unique in that regard, so it stood out from the crowd and grew a strong following.

Edit: The point is, it's not about how many people aren't interested in something, it's about how many people are interested in it. If a million people want something then it doesn't matter if ten thousand or ten million don't - either way you've got a potential audience of a million.

 

Exactly. I think a good example of this in practice is the asset store. I go in wanting to find a simple top down lighting kit, come out with 10 tabs open of all this stuff I didnt even know I wanted