Tuesday, March 1, 2011

Size isn't the only thing that matters: An analysis of open-world and sandbox games

Two major styles of game design which have begun to show up in increasing numbers over the last several years are open-world games, and sandbox games. There is still a good degree of confusion as to just what constitutes a sandbox game an what constitutes an open-world game, whether or not they may even be both the same thing, and exactly where the threshold between them lies. A large number of games, including big-budget titles, have experimented with sandbox and open-world design, but it's my opinion that few of them have actually leveraged the specific qualifies of those design approaches in truly effective and original ways. For me, the difference between the two doesn't lie in ambition, or narrative, or the size of the world, but rather, in the approach given to challenge and progression through the game. In this article, I'd like to take the time to deconstruct exactly what it is that defines an open-world game versus a sandbox game, using a number of examples, and the most important game design considerations for each of them, with respect to those aspects of challenge and progression.

Just semantics?

Traditionally, games focus on guiding the player either through a set path in a relatively closed or limited environment, mostly due to limitations of design, technology and narrative; it's just easier to create something and fine-tune it when you know precisely where the player is going to be coming from, going to, and what abilities they might have at that given point in the game. From a management perspective, it's also generally easier to coordinate development of limited but highly-detailed environments than it is to, say, construct a fully realised model of New York City, as seen in Grand Theft Auto IV. Open-world and sandbox games usually represent a much greater challenge simply due to their vast size, but I think coming to a clear idea of what separates open-world and sandbox games can help to determine the direction of a game's design, as well as focus on strengths while eliminating weaknesses.

On the surface, what makes these games might seem obvious: the sheer scale of the world and options available to the player. After all, most games of this nature in the past have featured immense scale when compared to more linear games. But looking a little deeper, that can't just be what it is. Although the size of the world is especially a consideration in an open-world game, to say that sheer size is what makes such a game is foolhardy. After all, FUEL had the largest 3D game world in history, yet its design utterly failed to capitalise on the scale the technology was able to provide, nor did it manage to motivate players to continue through the game after the initial “wow” factor. No, the differences between open-world games and sandbox games are much different, and are largely due to the fundamental differences in design approach.

Zelda resembles modern open-world games arguably more than modern open-world games themselves.
The Legend of Zelda is one of the earliest open-world games, even though it's largely not thought of one by many gamers. In Zelda, the overworld map is divided into a series of fixed “screens”, with brief environmental puzzles, mazes, or enemies standing in the player's way. These challenges serve to both inhibit the player's progress, but also to provide practice for the game's more significant challenges. These challenges come in the form of dungeons. Dungeons themselves are arranged like the overworld map, with multiple interconnected screens, but feature more maze-like layouts, more difficult enemies, and so on. Furthermore, when the player enters one, he or she is “locked in” to that particular area, and the deeper he or she moves into the dungeon, the greater the risk is, because leaving becomes more and more difficult, and more progress will be lost. At the same time, the rewards in dungeons are much greater, and to ultimately make progress in the game, the player is going to have to explore them fully.

The key thing to note, however, is that the challenge in Zelda's world is largely situated to specific instances, which the player does not have a significant amount of control over. While it is possible to gain new equipment, including more health, armour, and abilities that allow more damage, attacking at a distance, etc., these are all gained by plundering the depths of the dungeons. Sooner or later, the leisurely challenge of the overworld isn't going to be enough for the player, and the desire to explore distant and locked-off areas will continue to build. Thus, while the player has some control over how and when he or she experiences the challenges in Zelda, ultimately the player is still going to be experiencing them on the designer's terms, complete with all the challenge and risk that entails.

By contrast, sandbox games don't solely offer up this sort of localised, constructed challenge. In a game like Just Cause 2, the majority of the challenge comes from the game reacting to the player's actions. There are no specific screens which limit where the player can and can't go, and there are no set encounters which are significantly more difficult than the rest of the game. The “Heat” meter in Just Cause 2 indicates how much challenge the player has, and to a degree, the player has control over it. Cause a small amount of destruction, and the army might send a few soldiers to try to take the player out. If the player can deal with them quickly and effectively, their Heat level will drop. However, if the player continues on a path of violence and isn't able to dispatch of initial opposition, more challenging enemies will be sent, including attack helicopters, jet fighters, soldiers with heavy weapons, etc. As the player continues, greater Heat levels are unlocked, resulting in even more challenge. Heat, and therefore challenge, can occur pretty much anywhere in the world, so long as there's an enemy to take notice of the player's actions.

Got Chaos?  Just Cause 2 tracks player progress mostly per-explosion.
Progression in Just Cause 2 also takes perfect advantage of the sandbox format. Although there are missions in the game which are required to move the story forward, the bulk of play-time is actually dedicated to the random, wanton destruction that the player causes at his or her whim, managed by a “Chaos” ranking. The more Chaos the player causes, the more new weapons, missions, vehicles, and so on are unlocked. These, in turn, feed into yet more Chaos. Rather than tying the acquisition of new abilities and items to completion of set tasks, as in Zelda, Just Cause 2 rewards the player simply for exploring, experimenting, and causing as much destruction as possible. Since the game world is limited, the player is eventually going to run out of things to do, so in this case the size of the world exists not so much to give the player lots of options, as it would in an open-world game, but to effectively outlast the player's interest in the game. It's possible to “win” Just Cause 2, but it would likely take most players hundreds, if not thousands of hours to exhaust all of its gameplay, well beyond what most would likely dedicate.

Ultimately, then, what determines a sandbox game isn't how large its world is, but the way in which the player navigates through it, is presented obstacles to overcome, and makes progress, in whatever way it is measured. In a mere open-world game, the player may be able to explore at his or her leisure, but progress will still largely be governed by playing through specific designer-made challenges, be they dungeons, missions, etc. Sandbox games are, and should be, about what the player brings to the table, and providing fun mechanics that can be experimented with in fresh and interesting ways. That's why the distinction between sandbox and open world is so very important: the approaches taken in order to ensure successful design are completely different from each other, and why games like Grand Theft Auto and Red Faction: Guerrilla find themselves in a strange medium between offering up player freedom, and restricting them to a set path.

Designing for the open world

Open world games, in order to be successful, need to adopt a lot more structure than what many existing ones do in order to achieve cohesive design. The open world's strength is just that: sheer size, and freedom to explore and progress without too much restraint. Giving the player the opportunity to complete objectives in the way he or she wants to is what open world games should be all about. In Fallout, the player is given free reign of the world, with success in the game tied to solving a number of objectives. While its mystery-style storyline does push the player in certain directions, there is no set path through the game, and the player can actually leave large parts of the world unexplored if he or she doesn't want to bother with them, or simply comes across the ending on his or her own. Depending on the order the player solves (or doesn't solve) certain challenges, the game will respond accordingly. To me, this is true open-world gaming: give the player a task and set them loose on accomplishing it. Any game which attempts this, only to funnel the player through the game in a wholly designed order, has failed as an open-world game. After all, what's the point in a big, open world to explore and lots of fun rules if you still put the player through veritable linear corridors to move forward?

Many open-world games, unfortunately, don't know what to do with the vast world they've provided to the player. Some of them settle for allowing the player to do minor side-quests on the way to the main objectives, and others fill it up with extra abilities and bonuses for taking the time to explore. While these provide meaningful rewards to the player, they are wholly optional and secondary to progressing through the game. Sure, I could go and shoot all the pigeons in Grand Theft Auto IV for the modest rewards provided, but why bother other than for completion's sake? To provide a game with optional content is one thing, but to make the only motivation in completing that optional content an arbitrary percentage counter? That's just a crutch for lazy design.

Since open-world design is so tightly integrated with the way challenge and progress are presented, it's extremely important to provide the player with motivation for continuing in the game. At all times, the player must have an understanding of where he or she is in relation to the rest of the game, as far as objectives and story progress are concerned. Sub-goals are the easiest way to ensure this: the player has a final goal to work towards, but there are additional challenges along the way that must be completed in order to succeed. Furthermore, the player has to be given motivation, not just to complete that smaller goal, but that smaller goal must be related effectively to the progress in the rest of the game. Since different players are motivated by different things, this means the onus is on the designers to provide multiple types of motivations. Usually, this can be expressed in three ways: narrative, functional, and aspirational.
GTA IV's cutscenes aren't just there for flavour: they give you smaller goals and stories that motivate you to continue.
Narrative motivation is simple: provide the player with a good story reason for doing what they're doing. Obviously, this is contingent on a lot of things: the player has to care about the characters, places, events, etc. going on; the player has to have an understanding of the game world as a whole; the player has to understand the implications of success (and failure); and the player has understand why success or failure are contingent upon his or her action (best if it's not just a simple Game Over screen). Most of this can be handled in the early stages of the game, where exposition is even more necessary, but throughout the game it's good to remind the player why they should care about what's going on. Maybe this means a love interest being put in danger, the player's own abilities and power being threatened, revenge for someone already injured or killed... without these smaller events punctuating the larger quest, the player is going to lose sight of what's going on in the big picture as well. In an open-world game, story often takes a backseat to the player's freedom, and so it's even more necessary to make sure that the events are tied together strongly. Grand Theft Auto has traditionally done an excellent job of involving players in the story, with lots of interesting characters, twists and turns, etc. Each character on the map represents a tiny story in and of themselves, but also presents a smaller narrative arc, which in turn ties into the rest of the story as well.

Functional motivation is effectively what sorts of material benefit the player can receive from continuing through the game. This usually takes the form of a tangible reward, like a new item, a new ability, money, a health boost, and so on. Rewards are always best when anticipated and worked for, not simply handed to the player on a silver platter or seemingly at random, which means that it's always best to tell the player exactly what cool new benefit they'll receive if they keep playing. The important part of a functional motivation is that the mechanical benefit of the item is what is important to the player, not necessarily the effect the acquisition might have on the story, or its aesthetic appeal. Why should I care about getting the rocket launcher? Because it'll let me blow up those enemies that cause so much trouble for me! Why do I want the hookshot? I can use it to grapple around the environment! Functional motivation is effectively material, and ties into the player's empowerment fantasy, but it's no less important in keeping them going through the game.

Aspirational motivation is the most esoteric of these, as it deals with why the players are playing the game on a “higher level”. The aspirational motivation appeals to the sense of discovery, the player's vanity and aesthetic sensibilities, the desire to get all there is out of the game. Yeah, the hookshot might let me grapple to the dungeon's boss, but I'm really more concerned with how it will let me visit that cave I spotted off on the other side of that cliff! That new armour all well and good, but it's not important to me if I look silly in it – I want my character to be a badass! Aspirational motivations assume that players aren't just interested in one specific element of the game, i.e. the story or the mechanics, but are interested and fully invested in the experience of playing. The more such sensibilities can be played to, the better, and recognising how they interact with the narrative and functional elements of the game is also integral.

Moving through an open-world game should be an experience dictated by the player, but still controlled by the designer. The goal of an open-world game is to give the player a sense of freedom and control, but not to surrender it completely. An open-world game recognises that the player's interaction with the game is important, but that there still needs to be a strong supporting framework and specifically engineered challenges to test the player. Titles like Zelda, Grand Theft Auto IV, and Fallout all acknowledge this to varying degrees, and they are all equally successful.

Designing for the sandbox

If the strength of open-world games is in crafting a specific experience that the player is able to guide and manipulate on his or her own terms, then the strength of sandbox games is to provide the player with a set of tools to experiment with. In a sandbox game, objectives, progress, etc. aren't static things created by designers to be overcome by the player. As a result, sandbox games face quite a few more challenges than open-world games, since they can't rely on ushering the player forward with carrots in the same way more traditional games do. Sandbox games have to be fun, wholly and completely, on a base mechanical level in order to be successful.  Additionally, much of the challenge of the game has to come from the base mechanics and the way they interact, as well as the plyer, who in turn is also in control of game progress.  A number of design considerations have to be made in light of this.

Consider Roller Coaster Tycoon: while there are a number of specific rules the player has to follow, and a good deal of scenarios to play through, the game is effectively a toybox for the player to have fun with. Certain rides, including the eponymous roller coasters, can be hand-crafted by players to allow for a wide variety of outcomes, anywhere from gut-wrenching, to mildly entertaining, to fiery and horrific. The flow of attendees throughout the theme park is governed by simple AI, as well as their wants and needs at an individual level; this is all open to manipulation by the player, who is able to build gift shops, toll booths, walkways, food vendors, and much more. The game is much less driven by any actual objective than it is by the player simply fulfilling his or her own set goals, whether that's to build the prettiest amusement park, the most profitable, the most dangerous, the largest, etc. The way the attendees behave is directly tied into the different ways players choose to enjoy the game, meaning that they receive a challenge no matter what they try to do. Coupled with an easy-to-use interface, Roller Coaster Tycoon offers up a game whose mechanics carry it forward; the player does everything to make the experience meaningful.

Roller Coaster Tycoon is, despite its age, more a sandbox than most current titles.
In sandbox games, there are a number of ways to ensure that the player has a fun and challenging experience regardless of how he or she chooses to play. Tropico 3's “sandbox mode”, effectively an endless, objective-free play through the game, might offer one style of play, but within that, additional objectives can help provide flavour while still effectively maintaining a sandbox feel. Since the player's enjoyment of a game is contingent upon the success of those basic mechanics, it is of the utmost importance that those mechanics operate successfully and aren't prone to exploitation, bugs, and anything that would otherwise cripple the delicate balanced. Pared down to a few basic concepts, there are three governing factors which are most imperative in a sandbox game's design: action & reaction, constraints, and interdependence.

Action & reaction is the most fundamental, and while heavily important to all games, it's central to the success of a sandbox game. Put concisely, the player has to be able to exert some influence over the world, whether that is as a “god”, as a “commander” or as an individual character; then, the player must also be provided with accurate, easy-to-understand and clear feedback as to the effects of their actions on the game world. This applies to both simple and complex elements. In Just Cause 2, the player is informed instantly if his or her Heat level goes up, both by interface indicators and by a change in the pace, tempo and intensity of the in-game music, which begins to swell with dramatic strings and thumping bass. In Roller Coaster Tycoon, when the player increases prices for entry, he or she may see visitors leave or turn away from the amusement park gate.

In sandbox games, “doing things” is where most of the fun comes from; it's the joy gained from the mechanics themselves that motivates the player to keep going. However, if the player were capable of doing anything at any given time, not only might the game be over very quickly, but the player would have no reason to explore the game's systems fully. Constraints are necessary in order to focus the player in a direction of his or her choosing, while offering incentive to explore other aspects of the game that may normally go untouched. Roller Coaster Tycoon features money as its primary global constraint, but there are others, including the individual properties of rides and the sensibilities of the visitors at the amusement park. Note that a constraint is not the same as a goal: it provides motivation for continuing through the game and exploring new aspects of it, but it in itself is only a means towards another end, while actually dealing with a goal is usually what's fun in more traditional games.

Interdependence, is why a sandbox is any fun in the first place. A bunch of mechanics which are wholly independent from each other and have no influence might be fun to toy around with a little bit in isolation, but don't provide any meaningful decisions for the player to make and would quickly grow boring. Money in Roller Coaster Tycoon is a constraint, but it also allows for complex interaction between all of the different amusements and infrastructure the player can build: what happens if I spend all my money on drink stands, but don't make any bathrooms? How does that affect my park's visitors? Is it more worthwhile to charge lots of money at the door and less in the park, or vice-versa? In Just Cause 2, at the low level, the player has a grappling hook and parachute which may be used to scale the environment in extremely creative ways, but the ability to fight back at enemies is sacrificed for mobility, and poor navigation can lead to the player becoming an easy target for enemies. At a higher level, the player's successes in causing Chaos throughout the world open up new story missions, as well as new weapons and vehicles, which in turn can be used to cause even more Chaos – but the higher Heat level attained also means the player will be more vulnerable when using those new, destructive toys.

The key here is that while sandbox games can be big and open when it comes to geography and world size, they really don't have to be. “Sandbox” first and foremost refers to the ability to toy with the game mechanics in order to set new goals to overcome, and achieve creative and interesting effects. In the case of Just Cause 2, the huge map exists less more to exhaust the player than it does to provide a fully-furnished, extensive game experience as it would in an open-world game. A sandbox game can elect to have a story, and characters, and certain scenarios where the player may or may not be limited or driven by specific goals and constraints, but at a base level it has to be able to stand on its own as fun even when wholly divorced from the motivations that more traditional and open-world games require.

The hybrid?

Of course, games being what they are, there are many different titles which fall into a nebulous territory between sandbox and open-world. I've already mentioned Grand Theft Auto as an open-world game, which I stand by, but it also incorporates sandbox elements, namely, in the form of police officers and the Wanted level. Similar to Just Cause 2's Heat meter, if the player performs any illegal (usually violent) acts, the police will come after the player, first in limited numbers. These are fairly easy to escape from, but if the player chooses to fight back against the police, he or she will be beset by increasingly difficult SWAT officers, and even the military in some of the titles. While Grand Theft Auto isn't quite a sandbox game, it does have a level of challenge which the player is able to manipulate and isn't limited by scenarios the designers have created. The Civilization series could also be seen as a bit of a hybrid, with its potentially endless empire-building gameplay, but the strong competition and definite victory conditions means that it too doesn't quite qualify for the sandbox label.

Whether or not games should actively seek out sandbox-style gameplay when they have not necessarily been constructed as “pure sandbox” experiences in the first place is a difficult question to answer. I'm sure many gamers would appreciate a linear or open-world experience that has been very tightly refined, while others are more concerned with fun and don't mind if their games are unfocused, so long as they're enjoyable. For better or for worse, I don't think that this is a question that enters too much into the design of games: given how iterative development is for many, the “throw it against the wall and keep what sticks” approach can be beneficial, even if it results in something which doesn't quite have the unity and consistency of vision or elegance of something more precisely tailored.

No comments:

Post a Comment