Wednesday, March 31, 2010

This blog has moved


This blog is now located at http://inmachinam.decisionproblem.com/.
You will be automatically redirected in 30 seconds, or you may click here.

For feed subscribers, please update your feed subscriptions to
http://inmachinam.decisionproblem.com/feeds/posts/default.

Monday, January 18, 2010

Exploring the Breakspace



I’ve been playing a lot of Dragon Age: Origins recently and I’m still trying to figure out why I’m enjoying it. I’ve gotten far enough in the game to hit the point that I hit in nearly every 80+ hour RPG, the point at which I stop really caring about sidequests or character arcs and I’m basically holding down the escape key to skip past every conversation with a character not related to the main plot. The point at which I lose the patience to read about bandits robbing the town or the cave infested with spiders. So I must be playing because of the gameplay. Except I’m not really that interested in the battles – I feel like the important decisions I’ve made were choosing my talents in the level up screens, and the way I handle the battles has only a slight effect on the outcome. That said, there are some battles that are interesting, battles where that slight effect is the difference between winning and losing, but the vast majority of the battles are the standard slog through minions to get to the boss. The other main part of the gameplay, the character progression, is mostly a guessing game about the exact numbers behind text descriptions.

So what am I enjoying about Dragon Age? I think it’s this: there’s a system, and I want to break it.


Like most RPGs, Dragon Age is made up of several sets of rules which are distinct but still interconnected. The main ones in Dragon Age are battles, character progression, and conversation trees. Together, they form a complex system; the backbone of the game. Within this typical RPG system, there’s an interesting interplay between character progression and battles: the better you are at character progression, the less interesting the battles are, for the most part. That is, the stronger your character is the easier the battles are and the fewer difficult decisions you have to make in each battle, to the point where if you make really good characters the majority of the battles become trivial.

Which seems like it might be a bad thing at first. Why would you want to trivialize the main form of gameplay by succeeding at another form of gameplay? However, I think that a lot of the time, trying to make the battles trivial is interesting, and that process, not the battles themselves, is the core of most RPGs. After all, the ability to grind already trivializes most RPGs, basically making it so that the only obstacle between you and the end of the game is time. It’s not hard to beat an RPG – in the ones with random encounters, all you have to do is tape the dual stick forward, tape the X button down and leave the game on overnight.

So I’d argue that, instead of the battles themselves, the interesting part of an RPG like Dragon Age becomes breaking the system efficiently. The most interesting part is making the rest of the game less interesting. It’s fun to get Dragon Age on lockdown, as if it were a string of World of Warcraft instances you were trying to make simple so anyone could beat them. Maybe it even becomes the goal of the game: trying to break the system; trying to become so powerful that you just slice through bosses like a hot knife.

Which brings me to my point. I’d like to posit a new way of looking at “content,” or an entirely new type of content altogether. Instead of just the traditional form, I’d argue that there are two types of content – the first is the standard content we all know and love: the plot, the characters, the art, the number of encounters, everything that fills the space between the beginning screen and the game over screen and leads to a count of total hours that is often on the back of the box. Measured in standard content, Dragon Age is an 80 hour game, less if you don’t do the sidequests. Crackdown is 12 hours. Assassin’s Creed is 15.

But I think there is a second type of equally important content: the distance between learning the system and breaking the system – ‘breakspace’, for lack of a better word. Like the first kind of content, breakspace has a beginning and an end and it can be measured almost in hours. Like the first type of content, breakspace’s beginning and end are set in stone as soon as the game goes gold: your first shot and the optimal strategy.

So what am I talking about? What is the breakspace, and why is it interesting? It’s easier to explain by example and then come back to abstraction, so let’s talk about a simple form of breakspace, Tic Tac Toe. When you first play Tic Tac Toe it’s hard to understand the strategy and you end up moving more or less randomly. However, as you play more and start to think about it, you get better and better and move closer to the optimal strategy. Then, because Tic Tac Toe is pretty easy, you find the optimal strategy quickly – game over, you broke the system. That was the end of the breakspace in Tic Tac Toe.

In some ways, the breakspace of Tic Tac Toe is a lot like any other type of content. It has a beginning and an end. Like reading a book, part of the point of Tic Tac Toe is getting to the end and part of it is the journey that takes you there. You can spoil the end of Tic Tac Toe like you can spoil a good book – you can say: here’s the optimal strategy, it always ends in a draw. But a lot of the breakspace in Tic Tac Toe is understanding why that’s the optimal strategy, and going through the train of thought that led up to that optimal strategy, which is slightly different for everyone. Similarly, someone can spoil the ending of a great book without ruining the point of that book – a large part of the “point” of reading that book is not only knowing the ending, but appreciating how the author got there.

But not every game is as simple as Tic Tac Toe – most are incredibly complicated, and have optimal strategies that will probably never be truly understood. These are games like Chess, Go, Starcraft, Counter-Strike, etc. Because these games are more or less unsolvable, they have two forms of gameplay – the moment-to-moment decisions in the games and the breakspace, which is often called the “metagame.” However, the term metagame typically refers to the current realm of strategies and counter strategies that people use, whereas the breakspace of Starcraft is more like a story. The beginning is the first time you sit down to play Starcraft and fumble around with the units and the end is the unreachable optimal strategy.

So the sum “content” of Starcraft is both the high-level play itself and the journey towards the thought that creates these high-level games; the story about how we got to where we are now. Like most multiplayer games, some people start somewhere in the middle. Instead isolating themselves and taking the entire journey from the first game of Starcraft to where it is now, they read forums and jump right into the game a little bit behind its current level. They find value in Starcraft by seeking out high-level play, by trying to get to the frontier.

However, there are many people who do start at the beginning and get value out of a different part of the journey. They play offline with friends who are also just starting and so they climb up a whole different section than those who are aiming for the top. Because, ultimately, the value of Starcraft is in both the peak and the climb itself. It’s interesting to be at the top, but how you got to the top is also interesting. The people who start out online are at the peak, trying to get higher than anyone’s ever gone. The other people are way below, studying history – but in many ways, there is beauty in the history. Those who jump to the top do not experience the hundred strategies that failed, only the one that succeeded.

It is not only deep games like Starcraft and Chess, but every game, every set of rules with a goal which have breakspace. Many of these games have optimal strategies which will probably never be solved but which do exist theoretically for an infinitely powerful computer, e.g. Go, and their breakspaces are therefore theoretically finite but realistically infinite. However, there are plenty of games, like Tic Tac Toe, that have finite breakspaces.

One of these is Dragon Age. Dragon Age has tons of traditional content – a 60-80 hour plot, plenty of sidequests, deep characters, art, atmosphere, cutscenes etc. But this is only one variety of the content you’re experiencing when you play Dragon Age; the other is its breakspace. Like many games, Dragon Age’s breakspace is similar to its traditional content or a book or a movie. It has a beginning, the character creation screen, and an end, the optimal build or battle plan for every battle. It is likely very finite and if you sat down and solved it as a big math problem it would probably be significantly less than 60 hours worth of math (to get close; it’d probably be a lot more to prove that that was the optimal strategy). Like the game’s story, someone can spoil the end for you, and they often do. Go to any Dragon Age message board and you’ll see people jumping you ahead – “Dual wielding does better damage than two-handed,” “The game is easy if you get a lot of mages with crowd control,” etc etc. But even though you might know the end, it’s still fun to explore the breakspace, it’s fun to see why the other strategies are inferior, and it’s fun to see the story the breakspace tells. It’s a great story, created by a haphazard combination of developer intention and player discovery. You may know the end and the beginning, but the value of any story – the real content – is often how you got there.

[Crossposted at Game Design Advance]

Saturday, December 19, 2009

Single Player Character Progression


Recently, there's been a trend towards single player character progression, and not just in RPGs, but in action games like Dawn of War 2, Brutal Legend, and Assassin's Creed. As single player character progression gets more and more popular, I've started to think about what kind of gameplay single player character progression (hereon CP for the sake of everyone's sanity) creates. What kind of decisions does it give the player?

Why Character Progression is Interesting
Almost all CP decisions have a positive and permanent impact on some stat or your abilities. These are more or less the defining characteristics of a CP decision. For example, in Brutal Legend you can upgrade Eddie Riggs’ axe for an upfront cost and give it snazzy lighting damage, or fire damage. In Devil May Cry 3, you pay blood orbs to give Dante another move or upgrade his weapon damage.
In RPGs, CP decisions are usually very transparent about the exact details of their effects – 10% to spell damage, 20% to all fire spell damage, +100 mana, so on and so forth. For the most part, they give you the raw math that you need to understand which ones are good and which ones aren't so good.
So there’s a skill to deciphering the CP system in RPGs – you know the relative power of your fire spell to your ice spell, so you know that you want your character to take the fire damage increase. Deciphering the system an RPG lays out for you is usually a key part of the core gameplay. You're tinkering with the system; you're trying to find the combination of stats and abilities that the developers didn't quite balance, the math that produces the highest number.

Why Character Progression Doesn't Always Work

However, making CP decisions transparent in action games like Devil May Cry 3 or Brutal Legend is much harder to do, because there is so much more relevant information and that information is so much more obscure. When was the last time you ever saw an enemy’s health as a number in an action game? Or the numbers behind your damage? Typically, action games don’t show you the math behind the system.
So, logically, most action games simply don't tell you the math behind the CP decisions you're making either, because the fundamental system itself is opaque – you wouldn’t know what +10 damage was because you wouldn’t even know how much damage you deal regularly, much less the health of your opponents.

Consequently, CP decisions in action games are much more about luck than your ability to decipher the system. For stat boosts and items, this is definitely true whether the lighting axe does 5% extra damage or 10% extra damage could be the difference between it being worthless and it being completely overpowered, but it doesn’t tell you.  

It's even more luck-based when CP decisions involve learning new abilities in action games. Each ability has a dozen relevant stats – speed of animation, damage, how fast it moves you, whether or not it has invincibility frames, exact range, exactly what effect it has on you, etc etc. It's nearly impossible to tell which moves will end up being useful in these games just by looking at them on the CP decision screen. In some ways, action games are stuck here. Even if these stats were displayed, the player wouldn't understand what they actually represented – because action games have no intuitive way of giving the system to the player without filling the game with clutter.
Fighting games are similar to action games in this sense – they don't give you the math. However, when players get serious about playing, they spend a lot of time testing every single move and giving their results a number, and then comparing it to other moves and deciding whether it’s useful or useless. When serious players do this, they have an understanding of what the numbers mean – they created the number for health, so they know damage in the context of health, and they know how many frames a second the game runs at, so they know what an attack being 120 frames means. Range is less intuitive, but it’s often a reverse-engineered attribute, so they know what the difference between 700 and 600 range is because they made those numbers up to reflect their actual testing in the game. They reverse engineer the system themselves instead of the game giving it to them, so the numbers make sense to them because they created the context.
And it's this raw data that ultimately drives the players' final decisions. Every move in a fighting game is presumably “useful,” in the sense that there was a use intended for it, but only a few are actually worth using. That’s because it doesn’t matter that one attack is “long and slow” and one attack is “short and fast,” what really matter is exactly how long it is and exactly how slow it is.

The Problem of Permanence

Even in single player action games, players are often serious enough to reverse engineer the system in the same way – you'll often see numbers attached to gamefaqs guides on more hardcore action games. So why would reverse engineering the system work for moves in fighting games but not CP decisions in action games? The key difference is that CP decisions are permanent – so the process of figuring out the exact differences between attacks in order to make the right CP decision requires making the CP decision that allows you to test out the attacks in the first place. To contrast, you can really like a character's low kick in a fighting game, but realize that it's worthless against good players once you start getting better. At that point, you stop using it, no harm no foul.

However, that's not inherently bad. In most action games, it's a source of frustration – you're not going to redo 2 hours of a game once you realize that your initial decision to put points into the lighting axe was probably a poor one. But roguelikes have elegantly integrated the trial and error of figuring out the system into the core game itself.  In my experience, the reason why it works for roguelikes but doesn't work for action games is mostly in how they present it. In Brutal Legend, the game is content first, action gameplay second, and the character progression system as a nice little aside. You would never save and load over and over in Brutal Legend to figure out which axe is best because that's not the "point" of the game. But roguelikes make it very clear – the content is the system.

Friday, December 04, 2009

Ben Abraham's Permanent Death


You probably already know that Permanent Death is a iron-man (one death and you delete your save) run of Far Cry 2 done by Ben Abraham. I had heard about it a few months ago, but had only recently decided to read it after being reminded about it by today's Rock Paper Shotgun post on it. In a lot of ways, Permanent Death is incredible – it's incredibly well produced, it's very well written, and it has an absolutely fantastic concept. However, to me, it also misses the point of an iron man run in some ways, instead focusing on being (as it says on the cover) a "novelization" of Far Cry 2.

What I liked about Permanent Death

Permanent Death is really good in many ways. On the simplest level, it's always really interesting to see someone take a game and play it a different way than it's intended to be played, especially when the new ruleset dramatically changes the way you play the game, as iron-man rules often do. Also, Permanent Death has very high production values – the strength of the writing and presentation bind the whole thing together into a strong, coherent package.

One of my favorite things about Permanent Death, however, is that it's trying something interesting. Basically, it's saying – what happens if I play this game an entirely different way? How does the way I play it change? What's interesting about the new way I play the game, and how different is it from the way I play the original game, and what does this difference mean?

What I didn't like about Permanent Death

My biggest problem with Permanent Death, however, is that it doesn't really answer any of these questions. Instead of focusing on what's really interesting and different about an iron-man run of Far Cry 2, it focuses on a telling a narrative version of Far Cry 2's story with occasional interjections that remind you that it's an  iron-man run. It foregoes gameplay for characterization.

Let me give some examples. There's a lot of this in Permanent Death:

"Warren Clyde, was my saviour. He charged in, desert eagle and AK-47 blazing. He must have known I was going to pull something stupid, as he was obviously hanging around nearby. I’ll bet he came running as soon as the shooting started. How else can I explain way the almost divine timing that saw him turn up right at that very moment? A second later and it would have been lights out for Qurbani Singh. Talking to him later in the safe house, I barely remembered the rest of the journey. Apparently he picked me up, dragged me out from the middle of Pala by himself and put me back on my feet at the jetty on the north-western side of town. He even covered my retreat as I got in a boat and puttered downstream to collapse on the camp bed of the nearest safe-house. When I asked him about it, he simply gave me a thumbs up and said “Don’t mention it, man”."

And way too little of this:

"I’ve eaten enough lead to make a paperweight by now, but at least the engine’s running so I try and get to the front of the ship to see if I can’t take out the rocketeer. I didn’t even see the next rocket until the last second so I barely had time to duck out of the way. I hurt myself pretty bad in the process.”

As I read, I was hoping he'd explain his gameplay decisions – for example, why permadeath made him want to take an RPG over an IED, or vice-versa, or why he chose the gun he did, or why he approached from the angle he did. Battle descriptions are far too often "I ran into some guys and shot them and didn't die" and far too rarely about the actual moment-to-moment tactics of the battle.  Instead, there's a lot of exposition about how so-and-so made him go on some mission to retrieve such-and-such a thing, and what the guy looked like, and a whole lot of excellent writing about the character of Far Cry 2's protagonist and what he was thinking about.

Now, I'm not saying this is necessarily a bad thing. Permanent Death is an excellent novelization of Far Cry 2, and maybe that's all it's supposed to be. But it doesn't talk about any of the things that make an iron-man run interesting. It doesn't really capture how the game changes from normal to iron-man, and how his decisions changed. It's a good story, but as someone who's played Far Cry 2 there's just not that much there.

Tuesday, December 01, 2009

Goals in MW2: Other Thoughts


A column I wrote a week or so ago for GameSetWatch just went up here. I played a lot of MW2 before I wrote it, but I've played some more since and want to use this post to throw out some other ideas I've had and some things I didn't have room to put in the column.

1. I talked about how stat-tracking creates another goal (k/d ratio, mostly) that can conflict with winning each individual round. I still think that's true - however, there's definitely a lot of positive things about the stat-tracking and ultimately I think the game is better off with it in. For me, stat-tracking's main flaw is also the reason why it helps - it creates a new goal that is more individual than team based. The way that you jump on teams in MW2 feels random enough that the game would be a lot more frustrating if everything was focused on teamplay. Instead, having an average k/d is often what makes me want to play more - it's an individual goal that gives a rough measure of how well I'm playing and whether I'm getting better. To this end, it's much better than a win/loss ratio which is all over the place until you look at it averaging out over 100+ games, which means you have to play 100+ games to even start thinking about how good you are, and even then it's impossible to tell, just by looking at win/loss, whether you're getting better.

2. The advantage that higher-level players have over lower-level players in MW2 is much, much larger than the advantage they had in MW1, for several reasons. First, and most obviously, killstreak unlocks put lower levels at a huge disadvantage - chopper gunner and harrier airstrike are two incredibly important killstreaks to have in this game and you really can't get them both until level 20, and even then you don't have pave low or other important killstreaks until later. There was really nothing comparable to this in MW1. Second, the guns you got initially in MW1 were mostly considered to be the best guns in the game, so while your weaponry wasn't as versatile as other players', you weren't at a huge disadvantage. In MW2, however, the FAMAS and other starting guns are definitely strong, but I think they're much less obviously good than the M16 and MW1's starting weapons. Finally, pro perks! Every perk in MW2 has a "pro" form that you can complete challenges to unlock. The pro form of any perk is always better than the original form, sometimes significantly better. These are another high-level advantage that didn't exist in MW1.

As a result of these advantages, the MMO-levelling part of MW2 is a lot more important than it was in MW1 - one of the reasons why I think that the goal-dissonance it creates is a lot stronger than it was in MW1, where most people didn't care about it nearly as much. That, and titles. Probably mostly titles, actually. I want one of those pony ones so bad.

Sunday, November 29, 2009

Interface as Gameplay


Recently, I was reading an article by David Sirlin about minimizing unnecessary clicks when designing a game. One of the examples he used was the reload system in Resident Evil 5 – apparently, you have to click 5 times to do a fast reload of your gun (presumably like Gears of War) that you pretty much have to do if you want to be competitive in multiplayer. In the article, Sirlin declaimed against this "5-click tax" on the player every time he wanted to reload. Which got me thinking about interfaces in general.

Awhile ago, Blizzard announced that they were putting "multiple building selection" into Starcraft 2. A quick recap of the last thing I wrote on this blog: MBS is a change from the original game's interface that became highly contentious in the Starcraft community. Essentially, in Starcraft you can't select multiple buildings at once. In Starcraft 2, you will be able to. The main argument people use for this change is simple: it's an obvious interface improvement, because it streamlines the interface. Why wouldn't you put it in?

I'm not going to argue that Sirlin's wrong in criticizing Resident Evil 5's reload system (I've never even played the game) nor am I going to talk about the pros and cons of multiple building selection here (more on MBS here). Instead, I want to contest the implicit assumption behind these two arguments, and behind most discussions about interfaces in games: that there is a difference between a game's gameplay (the real game) and its interface (not the real game).



A screencap from a youtube-demonstration of the quick reload.


A screencap from a youtube tutorial of quick reloading in RE5. It's actually kind of a cool emergent mechanic (or exploit, depending on how you look at things).



I think this assumption is a bit easier to see in the multiple building selection debate in Starcraft 2. When people say that multiple building selection is an obvious interface improvement, they're implicitly dividing what a player actually does in Starcraft into two broad categories: the "real game" and the hoops you have to jump through to play the "real game" – which we'll broadly label as the interface. If you talk to people about strategy games, especially real time strategy games, you start to notice this distinction a lot, implicitly or explicitly. There seems to be a very large majority belief that the "real game" of Starcraft is the part where you're basically the commander-in-chief – you're making split second decisions, ordering flanks, deciding when to expand, deciding where to place your army. Unfortunately, though, because humans cannot communicate telepathically with computers, this clumsy thing called the interface had to be created in order to play the "real game." In this argument, the interface is basically an unlucky necessity – we have to use it, so let's make it as intuitive and streamlined as possible, and get it out of the way so we can get back to playing the "real game." Sirlin's article has sort of the same implicit assumptions, although on a smaller and more arbitrary scale. Obviously, aiming, which is obviously just about handling the interface, is part of the "real game" in FPS games, but clicking to reload is not.

On a basic level, this line of reasoning does make a lot of sense. People want to create clever strategies in Starcraft – a strategy game! – and losing at the game because the other person could handle the clunky interface more adeptly is frustrating. However, if you look a bit more closely, it falls apart. If there's a difference between what is purely interface and should be "streamlined" or "improved" and what is gameplay, what exactly is it? How do you tell which one is which? Where does one begin and the other end? In my mind, the answer is: you don't, because there is no clear difference. In fact, there often is no difference.

At this point, I'd like to differentiate between interface that obviously doesn't affect gameplay and interface that could possibly affect gameplay. Let's arbitrarily call the former harmless interface, and keep the latter as interface in general. There's a reason for this: I want to avoid confusing selection mechanics with harmless interface elements that can obviously be streamlined without changing anything. In most games (although not all), this means menu screens, high score boards etc.

In most games, we can break gameplay up into two types: decisions and actions. In Starcraft, decisions are, well, decisions – should I tech up or rush him? Should I move my units across this bridge? Can I chase his fleeing army, or can he punish me for it? Should I attack his expansion or his main base? – while actions are more about how quickly and efficiently you can click. For the most part, decisions are mental, while actions are physical. To most casual Starcraft players, decisions are the "gameplay" the part where you get to pretend to be a general, and actions are the ugly externalities of the "interface." But are actions in Starcraft entirely about the "interface?" What about split-second decisions, and what about proper micromanagement? Is there a difference between being able to move your units to flank quickly and being able to select more than one building?



It's hard to find SC2 screenshots with the interface, so why not just skip it altogether?

I tried for an hour to find a screenshot of the SC2 interface that wasn't grainy youtube quality and couldn't find a single one, so here's a battlecruiser in space.



Let's leave Starcraft for a moment and take the distinction into another game: Guitar Hero. In Guitar Hero, the distinction between gameplay and interface is absurd. But within a song itself, the interface – the buttons on the guitar that you press – is the gameplay. If we made the distinction that Sirlin's article makes between reloading and the rest of Resident Evil 5, or the distinction that Starcraft players often make between mechanics and decisions, the only "real" gameplay in Guitar Hero would be deciding when to use star power – the only decision you actually make at any point in the game. The rest of the game is totally mechanics - there is absolutely no line between interface and gameplay. Guitar Heros' interface is the controller, so your ability to interact with the controller is your ability to use the interface, which is basically all of Guitar Hero's gameplay, minus star power (which you have to waggle your guitar to get going, so there's even a bit of interface in that too). While the distinction between interface and gameplay in Starcraft is somewhat intuitive, the same distinction in Guitar Hero makes no sense whatsoever – Guitar Hero is obviously a game about controlling the interface.

Now we can look at a slightly less obvious example – Resident Evil 5. Resident Evil 5 obviously has considerably more decisions-based gameplay than Guitar Hero, but still a significant amount of actions-based gameplay – most notably, aiming and shooting, one of the cornerstone mechanics of the entire FPS genre. But what's different about paying a "5-click tax" to quick reload and having to aim your gun every single time you want to shoot someone? Both are purely mechanical like clicking Guitar Hero buttons, so why is one considered an externality of the interface and the other a respectable part of the "real game?" Why isn't aiming a 1.5-inches-of-hand-movement tax?

The same applies to a real-time strategy game like Starcraft. Where does one make the arbitrary distinction between certain types of interface-interaction – fast-paced micromanagement or the ability to accurately and quickly move the mouse – and others – the amount of units you can select or the speed at which you can click hotkeys? Why is the former considered part of the legitimate, "real" game while the later is an externality of the interface that should streamlined and removed if possible? As far as I can see, there are no reasons for these distinctions. They are born, presumably, from the difference between what the game is, and some semi-communal idea of what the game should be like. But in the reality of the game, the "interface" is as much part of the mechanical gameplay as anything else. There is no practical difference between having to click 5 times to reload, having to click 5 times to select buildings, and having to hit 5 notes quickly to beat a song in Guitar Hero – there is no reason one should be considered a result of the "interface" and the other a result of "gameplay."

Common counter-argument: An RTS is a strategy game, therefore the game is intended to be about strategy, not about fast clicking. One flaw in this argument: if the game was intended to be about strategy exclusively – that is, decisions and not mechanics – it would not have been an RTS, it would have been a turn-based game. In any RTS that is played at a competitive level, it will always be better to be able to click faster. But an even deeper flaw is that this argument implies that a game should be played how the developers intended and gameplay should be restrained if it threatens to take a new course – which essentially precludes any possibility of emerging, interesting mechanics and leashes the game to being solely about the developer's original intentions, and not a collaborative process – a conversation – between developers and players.

That said, I don't think that removing single building selection in Starcraft is bad, nor do I disagree with taking out the 5-click quick reload in Resident Evil 5. This also isn't a rant against streamlining interface – in general, I think making the interface part of gameplay more intuitive can be very helpful. What I'm arguing here is not that these things are good, but that if they are bad it is because they are boring, unintuitive, clunky, or otherwise bad gameplay mechanics. In other words, the inclusion of multiple building selection might be good because it creates interesting gameplay or gives the game a better flow. The 5-click reload should be taken out because it is tedious gameplay. Interface should not be considered excess fat to be trimmed off the "real" game.

Saturday, March 22, 2008

Multiple Building Selection in Starcraft 2


Multiple Building Selection (or MBS) is a rather simple UI feature that exists in almost every recent RTS game. In many old RTSes, the player can only select one building at a time – so, in a game with a lot of production buildings, the player must either devote all his hotkeys to building selection (leaving none for his army) or all his time to jumping back to his base and quickly selecting his buildings and tapping unit production hotkeys. At high levels of play in Starcraft, most players play a careful balancing game with time – at some points during large battles, one must jump back to their base and produce units, or else one will fall behind in army production and simply be overwhelmed, no matter how well one can dance his units. However, one must also pay attention to micromanagement of his units during crucial battles, or else the opponent will be able to skirt around his army and decimate a much larger force with several well-managed units. This conflict – how much time one can spend on macromanagement and how much time one can spend on micromanagement without falling behind in one or the other – is the fundamental principle of modern competitive Starcraft. Because it is physically impossible to accomplish both perfectly, i.e. to macro with 100% efficiency and micro your army quickly and cleverly at the same time, players are constantly striving to improve at both and find a balance between the two that allows them to gain an edge.

Because this balance is a delicate and crucial one, MBS is pretty controversial in any serious discussion about competitive Starcraft and its sequel. The ultimate question is: will MBS affect this balance so much that competitive Starcraft 2 will suffer, or is it ultimately a largely inconsequential change that will really only show a significant effect in low-level play?

Streamlining the UI is not Always Good.

When discussing MBS, a lot of different arguments get tossed around. Some people say that MBS is simply streamlines the UI, and therefore allows the game to be more accessible and less frustrating. These people usually go on to say that the UI is really just an artificial barrier between what we want to happen and what we can tell a computer to do using a keyboard.






Firstly, making the UI easier is not always good by any means. Part of any game is the way we interact with it, and what commands it will understand. It’s quite obvious that changing the UI in an FPS will affect how people play that FPS – if Counter-Strike had an auto-aim feature that always got headshots, it would clearly be a different game. Moreover, adding an auto-aim feature is clearly a UI change – the mouse is simply the tool we use to aim, as the keyboard is the tool we use to select buildings, and just like we would always like to select all our buildings in an RTS, we would also always like to get headshots in Counter-Strike. Take, for a more concrete example, the Decursive controversy in World of Warcraft. Decursive was a simply UI modification that allowed the Priest to quickly and efficiently de-curse allies during a hectic encounter. Because the ability to remove curses was crucial, this UI mod changed Priest gameplay significantly, and made raid instances much easier. Decursive didn’t just make the game easier for new players; it changed World of Warcraft’s gameplay on a fundamental level, and made it a good deal easier to play.

Starcraft is an RTS!

Some people say: well, yes, changing the UI will affect Starcraft 2’s gameplay. But Starcraft 2 is an RTS, not an FPS, so strategy should be the thing that determines who wins and who loses, not how many times you can click every minute!

Well, no. Yes, RTS has the word strategy in it, but there is one key difference between a TBS and an RTS: an RTS is real-time. In a turn-based game like Civilization and Dominions, streamlining the UI is almost uniformly welcome, as those games truly are meant to be about strategy and very little else. However, an RTS is about clicking speed, in the same way that an FPS is about how well you can aim; that’s what makes Starcraft different from Civilization. In both RTSes and FPSes, strategy and tactics are important, but so is one’s ability to move the mouse, and moving the mouse is an important and interesting skill. Guitar Hero is a game purely about skill, there is no strategy, yet Guitar Hero is interesting and difficult to learn and master, not simply an object of rote memorization. The same is true with DDR, Beatmania, or even the act of playing a song on the guitar. All of those things are skills, in the same way that being able to use the mouse quickly and efficiently is a skill, yet none of them are trivial or any less impressive for being “just a skill.”





Starcraft is an RTS, not a TBS. It is truly difficult and interesting to gain APM (actions per minute) and increase one’s speed, and that has, as a whole, helped the conflict between micromanagement and macromanagement shape Starcraft as a competitive game. Clearly, strategy matters – but it’s not the only thing that matters, nor should it be.

Depth in Competitive Starcraft and Starcraft 2

So far, two things have been established. One, streamlining the UI does affect the gameplay of an RTS. Two, the skill of managing the UI and one’s ability to use the mouse is an important thing that has helped to shape the unprecedented success of competitive Starcraft. Therefore, the issue of MBS must be taken into consideration – does it actually trivialize the skill of APM and the divide in attention, or does it simply redirect its focus? One of the skills that Starcraft’s current pro scene rests on is the division of attention between micromanagement and macromanagement. However, the fact that the division is between micromanagement and macromanagement is less important than the simple fact that there is a division at all. There is a basic idea that drives competitive Starcraft forward: in order to play perfectly, a player should really be doing five separate actions at once – however, that is physically impossible, so the player is always trying to react more quickly and weigh the tasks so he spends less time on tasks that are less important. In this way, competitive Starcraft has evolved from its early micromanagement-heavy stages, through its middle macromangement-heavy stages, and now to its current state, in which the two are weighed somewhat equally.

However, it’s not really important that the division is between macromanagement and micromanagement. What’s important is that there are always several actions which the player wants to do at once, but he can only focus on one. These decisions, as long as they are interesting and important, could be all about micromanagement or all about macromanagement.

In recent demos of Starcraft 2, Teamliquid.net fielded two good players, who, although nowhere near pro-level, reported to be able to macromanage almost perfectly in Starcraft 2, i.e. there was never a time when they wanted to build units and had the money but weren’t able to because of micromanagement in a large battle or paying attention to the timing of an expansion, and they attributed this largely to MBS. During these debates over MBS, many people tend to use Warcraft 3 as an example of a game with trivialized macromangement that still worked on a competitive level. However, Starcraft 2 is shaping up to be more like Starcraft and less like Warcraft 3. Competitive Warcraft 3 is interesting because the decisions in micromanagement are interesting, because there are so few units and every one of them counts. However, in Starcraft (and therefore probably Starcraft 2) units are more expendable, are produced more quickly, and die more quickly – therefore, micromanagement is less crucial than it is in Warcraft 3. Will the micromanagement decisions in Starcraft 2 really be interesting enough to balance out the relative removal of macromanagement through MBS? It’s hard to say without a physical copy of the game, but if Starcraft 2 wants to be anywhere near as deep (and, therefore, as competitively successful) as its predecessor, it’s going to have to make up lost ground. MBS can only narrow the focus of the game; it can only remove important decisions and conflicts, and Starcraft 2 is going to have to struggle to find a conflict in attention and interests as important as the one that gave Starcraft such incredible depth.