Wolfire Games Forums

SuXoR wrote:It also unlocks the hellhole of combat and physics glitch abuses that will dominate multiplayer games. All of the exploits of the engine, environment, and combat systems will likely make multiplayer very annoying unless wolfire constantly updates and patches the game. Also servers..

Pretty much. If there's an edge or dominant strategy to be found and exploited, and there nearly always is, then rest assured it will be found and exploited. People are smart like that.

And this can also have the side-effect of fundamentally skewing the way the game is played - partly because of the natural tendency to exploit the system to gain an advantage, but also because human opponents require different tactics to computer opponents. If, for instance, close combat was very twitchy, and people discovered that they win more often by throwing knives at their opponent than engaging in close combat, with or without weapons, then the feel of the gameplay would be quite different from the single-player... at least until the patch which nerfs throwing knives but leaves wolves a bit too powerful.

Mount&Blade is a good case in point for this - designed initially as a single-player game, until the engine was revised to add multiplayer in Warband. I had always wanted to see this in Mount&Blade - I thought the complexity of fighting a smart opponent would take the game to another level, where teams could use intelligent tactics to gain the upper hand on the battlefield - but after fighting against other players online I no longer recognised the game I had so enjoyed playing alone: the reality of the implementation did not align with (admittedly naive) expectations. Not that the strategies the players devised were stupid (I doubt I would have discovered them myself if they weren't already used ubiquitously), but they simply belonged to a level of metagaming that I wasn't interested in.

I probably wasn't alone. There were servers dedicated to dueling, for instance, where people were expected to use melee weapons only and had to wait to engage opponents 1 vs. 1. The game didn't enforce this as far as I am aware, just people who didn't want to have to play as an archer or knight to compete with other players, and although it eschewed the popular way of resolving mass melees by skewering people from behind there was still a large emphasis on who could make contact with their weapon the quickest (by moving and turning in a certain direction whilst swinging your weapon). Imagine the leg sweep in Overgrowth coming down to that, positioning your avatar in such a way so that the game registers your attack closer to the beginning of its sweep than your opponents.

In an attempt to balance the game, I think the developers also introduced numerous mechanics such as kicking, breaking blocks with heavy weapons, becoming unbalanced if swinging with certain weapons, and changing stances with spears to better combat cavalry. Personally, it just seemed to make things more convoluted because the game wasn't designed with multiplayer in mind to begin with, and no amount of extra abilities is going to achieve that balance. (Course, you could argue that even games designed specifically to be played competitively between multiple people are never perfectly balanced, but you stand a better chance of achieving a happy balance if you allow it to influence the way the game is designed early on.

This essentially is the problem with adding the kind of competitive multiplayer in Overgrowth that the OP is asking for. To work well it has to be planned and designed as part of the game from an early stage (like Starcraft or Streetfighter), and not tacked on as an afterthought or a novelty, which entertains people for a short while but ultimately lacks the depth/richness of gameplay to keep people competing for a long time.

For example, if maps which are currently designed to be played in single player were to be used for multiplayer they would also need to be designed to keep the movement of players flowing around the map and not constantly gravitating towards the same small area (otherwise you'll get bored of it quickly), and you also have to give places strategic weakness to maintain balance (so no ledges and towers only accessible by climbing where someone could simply wait and kick other players off as they pull themselves up). Then with the gameplay itself, you have to prevent any one element (weapon/attack/creature) from being dominant, usually by introducing as many levels of strategic depth as possible, encouraging players to out-think their opponents and anticipate their actions (I think the word "Yomi" in Japanese means something to this effect). So you have your first level of depth, the basic attack, but then there is a second level which you can use to counter the basic attack, and if you guess that your opponent will open with the second level attack there should be something you can do to counter that. And so on. The more levels the better. The best multiplayer games have lots of layers, and also allows the inexperienced player to occasionally win, e.g. when the experienced player is so used to countering the second level or third level of opening attack, and so doesn't expect his opponent to open with the most obvious first level of attack (which thus is itself a counter of higher levels of attack and bring things full circle to ensure the mindgames never stop). Imagine doing that for each form of attack, each weapon, each creature, and even locations or routes in maps if you plan on CTF.

Endoperez already brought up financial and time-related reasons why this specific kind of multiplayer - the hardest kind - might not represent a worthwhile investment to the Wolfire team, especially when one of their aims is to capture the experience of cinematic combat (which would be even harder to achieve in multiplayer given the reasons I stated above). That's not to say that no kind of multiplayer could work, or that I wouldn't want to see competitive multiplayer if it could be done well - it's just, with a team of this size (talented though it is), it's a very big ask.

That was a very, very nice first post. The fact that it mentions me by name and happens to support my personal opinions might have a little to do with this... : P

Welcome aboard!

I agree, and I really hope this game gets some mp treatment. It's be PERFECT for multiplayer! I imagine it's be like Jedi Academy, where everybody would have their own unique fighting styles. That would be way too cool.

Endoperez wrote:That was a very, very nice first post. The fact that it mentions me by name and happens to support my personal opinions might have a little to do with this... : P

Welcome aboard!

Thanks, Endoperez.

To be fair I wrote a little more than I intended, and in reflection probably sound quite dismissive of Polyoptics' desire for multiplayer fighting. I completely understand why he'd want multiplayer in Overgrowth; the idea of a fair competition between players (rather than an AI that might cheat in order to balance the challenge), solving the tactical problems that intelligent opponents would provide, and forging strong social bonds within a community through play, etc. is all very appealing. I think the norm is that we play together, not by ourselves, and as technology advances this will probably become the main way computer games are enjoyed in the future.

Fein wrote:I agree, and I really hope this game gets some mp treatment. It's be PERFECT for multiplayer! I imagine it's be like Jedi Academy, where everybody would have their own unique fighting styles. That would be way too cool.

Yes, and weapon preferences to match etc.. If done with the same care that has been demonstrated by the dev team so far, it could be very interesting.


At this point in development, IMHO, i feel my time is better spent to think about why it could be great and how it might work. Not to dismiss some great posts in here, it just seems a bit early for a "bah humbug" approach.

I suppose I would rather leave the details to Wolfire, as they have created what we see so far, it would only seem to make sense that they are the only ones qualified to diagnose the issue at that level.

However, it is interesting to think about some of the pitfalls and difficulties the OG team will face.

I've been following this game for a while, and I have to say it looks really interesting. One of the things I looked for was multiplayer, so I thought I might add a bit to this thread. That being said, I'd totally play the singleplayer too.

I basically agree with what illogicalAlready said. IF multiplayer (a good and fairly balanced one) was easy to add - then I'd love that. But you can't really just slap on something like that.
I'm a Mount & Blade player too. When the first game was out, the idea of multiplayer was there, but pretty much everyone thought it would be impossible - lag that impacts reaction times, ping, a small developer team etc.
But then the company released an expansion pack - and it had multiplayer. In my opinion, it's actually working really well. There's a competitive scene with leagues and what not. Of course lots and lots of time had to be spent on balancing things out, though.
About the duelling part: those things you mentioned (and more) do exist, but why would you view that as negative? It contributes a great amount to making the game enjoyable to compete in. It's more of pushing the system to its limits and discovering new ways of beating other players. The problem newer players have is that is simply looks.. confusing. But this could be helped by better animations - not by dumbing the game down.

I think a good idea if multiplayer ever became part of any plan, is to focus attention on it like an expansion pack. Because that's what is needed if multiplayer is ever going to be any good: focusing attention on balancing and building things around that. Which is a very time consuming project, especially for a small dev team.

Yes, I agree that multiplayer should be an expansion sort of thing. I don't know if they would charge for it or if it'd just be a free DLC, though. But that's what I mean: They make OG's single player. Then, if they still want to, they can put the effort that multiplayer needs into it. Singleplayer comes out great, multiplayer can get it's fair share of detail and work too without making everybody wait for singleplayer. But they do have multiplayer in the 'plans'. It's more likely to be co-op, and if something needs to be cut from development then multiplayer is higher on that list since it's non-crucial to the actual game (At least that's why I'd suppose it's higher.)

The alpha updates are showing a lot of changes in the combat system, IMHO, thats the best time to figure out something that will work for multiplayer.

Leaving it to an expansion pack later on may mean that the whole combat system might have to be reworked, or worse, they must squeeze multiplayer into a system that was created without any regard to it, leading to a 'hacked in' experience.

Even though it may or may not delay the game, I think it goes without saying: The sooner multiplayer mechanics are being thought of and considered, in the development cycle , the greater the chance that the end result will be a cohesive and solid experience.

No, it doesn't have to look hacked in. Hacked in is when you go into Versus mode on Turok 2 and use level warp to bring a friend to the main game. It doesn't work properly and is obviously unintended.
And yeah, working multiplayer into the game before release WOULD delay final release, by quite a bit in fact. That's because it takes time and effort to remove the hacked in feeling.
You don't have to rewrite combat if you leave multiplayer for an expansion pack later, especially with OG because it allows any character to be a player. The only real changes are setting up a client-server interaction that ensures that all players see the same thing at the same time. This involves hosting the game, the assets and the other bits of code and things. So you rewrite how the code accesses a lot of things, but the base mechanics remain the same. If David's code is worth its salt then it shouldn't be an impossible process to add multiplayer after the main game is finished. I'm pretty sure David knows what he's doing, so whenever/if they start developing multiplayer, they'll do it right. You have to keep in mind that this is a product they're producing, not some fan made hack/mod that doesn't have to be professional. (Not to imply that fan made mods aren't great)

Hubbel wrote:About the duelling part: those things you mentioned (and more) do exist, but why would you view that as negative? It contributes a great amount to making the game enjoyable to compete in. It's more of pushing the system to its limits and discovering new ways of beating other players. The problem newer players have is that is simply looks.. confusing. But this could be helped by better animations - not by dumbing the game down.

Mainly because it felt a little obtuse. I could block manually, and I could feint (although this was changed at some point), but they weren't much use if you didn't know how to attack effectively. I agree with you about animations helping; it wasn't intuitive because you didn't always get clear visual feedback to understand why you were failing or whether lag was affecting your attacks, feinting and blocking. But it does highlight a problem if you are a developer planning to introduce a multiplayer component further down the line - you don't want to be making new animations months later and so forth.

Also, the skilful attacks in M&B, while not bad design per se, are not in the game because the developers sat down and planned it that way. The players pushed the system in a way the developers probably hadn't thought about too much and then they sought to integrate it within the design of the whole game and control how players use it.

It pretty much boils down to what Polyoptics said about planning early so when you implement multiplayer later it will go more smoothly, and not change the feel or nature of the game inadvertently.

Agreed. Multiplayer implementation is very complicated and a unique combat system like OG will not make it easier. IMO, the only way to get it right would be to design the combat system with multiplayer clearly in mind.

Adding it in later would be a lot more work. If the combat mechanics had to be significantly altered (which would be extremely likely) to account for balance (which is of much lesser importance when dealing only with ai) it would be an uncomfortable experience for anyone who was already playing the game.

Yes it would of course delay the release of the game, which is fine by me, but probably not fine by some players here. That's Wolfire's call to make.

I can imagine walking into the dueling arena, up to a weapon rack and proceeding into an intense environment of combat and acrobatics. Fighting to the death. There is a heap of potential there!

That's one thing: If multiplayer is going to limit the combat system severely for balance issues, it shouldn't be implemented. The combat needs to be balanced as is, but because it is skill based, not level based, balancing will be a much smaller problem for OG. But the combat wouldn't have to be rewritten or redesigned. That's the whole point of spending a good amount of time working on and polishing multiplayer: This way things don't have to be limited, changed or ruined. And the combat doesn't have to be written with multpilayer in mind because multiplayer kind of is something you tack on to the end. Unless it's a game like Brink, you don't really have to always have multiplayer in mind, especially if co-op is much more likely than PvP.
As for the arena, I see issues with weapons being used there as the weapons make fights go extremely quickly. The fist fights are already exceptionally brief and adding swords just ends them faster unless the two players are both very good at parrying. So a one on one duel is going to be someone messing up a block, being disarmed and then sliced to pieces. That or someone gets back kicked into a wall and breaks their neck. The fights are too fast and scrappy for a one on one in an open arena. The jungle gyms that Aubrey might help, but you still run into the fact that OG's combat is not stylized, is not flashy and exaggerated and is not slow and easy to follow.
That's what I noticed about Mount & Blade: The combat is very slow feeling, even when turned up faster. The sword fights are very static. Block, swing, get blocked, block. The mounted combat is where M&B really shines, which is and isn't a problem. I love the mounted play because it feels more dynamic. It's also more challenging when you face another mounted unit because you have to work harder to line the arrow up, or position yourself for that strike. But it lacks that speed and ferocity you get from Overgrowth's melee. But the speed that fights are engaged, raged and won is just too quick for one-on-one. It's not bad for co-op.
In fact, you can kind of do co-op already, just with the AI. The fox character I made is set to Turner's side, so if you spawn a fox Turner they will fight any enemies they find, but not each other. The only thing it needs for buddy AI is the ability to follow a player and to follow basic commands. I'm not sure if David was thinking about buddy AI, but I love to see that sort of thing when it's well-executed.

I have to disagree with you, Korban. In single-player, the AI reacts in certain ways. Attacks can be balanced in relation to how well the AI can react to them. If the AI gets better at blocking an attack that's been done several times before, spamming single attack becomes an poor single-player mechanic. However, what if the AI becomes better at blocking that single attack than what most human player's can hope to achieve? If it's like that, in multiplayer, there would suddenly be no downside to spamming an attack.

The only way to change that is to change the attack(s) that are spammable so that they aren't as easy to make repeatedly. Or, those attacks would have to be rewritten or redesigned.

This can happen for anything that an AI and a player react to differently. Which is pretty much everything.

Or you can see that your AI gets to good, so you dumb the AI down to an ability more synonymous with a person's. That's what it feels like David is going for with his AI: As close to another player as possible. He's doing really well in that the enemies can do every attack the player can, pick up weapons, run for help. The only things left are parkour and improving what exists really. It's obvious that the AI can become to good, we see it all the time. That's why you limit how smart the AI can get. That doesn't limit the programmer's freedom, it just limits the AI's competence.

Korban3 wrote:Or you can see that your AI gets to good, so you dumb the AI down to an ability more synonymous with a person's. That's what it feels like David is going for with his AI: As close to another player as possible. He's doing really well in that the enemies can do every attack the player can, pick up weapons, run for help. The only things left are parkour and improving what exists really. It's obvious that the AI can become to good, we see it all the time. That's why you limit how smart the AI can get. That doesn't limit the programmer's freedom, it just limits the AI's competence.

That's like saying "I need to make a building that can stand an earthquake, so let's just make the buildings invincible. Diamond is the hardest thing known to man, so let's make it from diamond."
That is, the goal isn't bad (make AI which can challenge a human), but you can't do it the way you suggest (make the AI play as a human would). Not only that, but you're ignoring that it's impossible to do it (no house is invincible, no AI truly equals a human), and it wouldn't be as good as you think it would be even if the goal would really be reached (the diamond shatters, playing against competitive humans isn't the same kind of fun as current OG combat).

The AI should be made to challenge a human player in a way that's fun and makes for a cinematic fight.

A human would prefer to challenge another human player in a way that's unfair and makes for a one-sided fight. The other player, if he's good, would adapt. They'd probably end up not walking or running much, but instead rolling and jumping in weird ways, trying to confuse and out-guess the opponent. That would make for a horrible AI. Especially if it really could beat a player that way.