Gamasutra: The Art & Business of Making Gamesspacer
View All     RSS
October 25, 2014
arrowPress Releases
October 25, 2014
PR Newswire
View All

If you enjoy reading this site, you might also want to check out these UBM Tech sites:

EVE Online and the meaning of 'sandbox' Exclusive
 EVE Online  and the meaning of 'sandbox'
October 29, 2012 | By Kris Graft

"Sandbox gameplay" is one of those ambiguous terms in game development that is used a lot, but rarely defined.

To Matthew Woodward, senior designer at EVE Online developer CCP, "sandbox" is just a mindset in which his studio operates on a daily basis.

"['Sandbox'] is really hard to define, because it's so ingrained in what we do, that we know what it is, but it's hard to put into words," he tells Gamasutra.

But he offers up a pretty good definition anyhow. "Basically it's three things: being social, goal-driven and emergent -- making a game open, giving players control, essentially, and if you're making a multiplayer game, making it as social as possible, because that's why they're playing the game in the first place."

Woodward posits that emergence is likely the most important aspect, and it's something to concentrate on whether you're making a single-player sandbox game or a multiplayer game.

CCP is a studio that knows a thing or two about open-ended MMOs. EVE Online is nearly 10 years old now, and its trademarks are exactly the three elements that Woodward uses to describe sandbox games. EVE's player-driven economy, professions and close ties with the community have all helped the game standout from World of Warcraft copies, and carve a successful niche of its own.

eve online.jpgKey to EVE's success is grasping what it means for an MMO to be self-sustaining. Woodward explains CCP's approach. "Essentially, you need the emergence, you need the openness.

"The big enemy is rest states -- a place where players keep on doing the same things over and over again. That's the big thing that's going to drag down your open-endedness. That and obviously if a player can finish [the game], it's not open-ended," he says."

That open-endedness can translate into longevity for an online game. "[The game has to enable] an ongoing balance. Anything that supports competitive-type gameplay, trying to be the best at something, is really good for open-endedness, because you can have that back-and-forth between players, and you always try to get it back when you lose it," says Woodward.

Narrative vs. sandbox

There's one element that is not mentioned in Woodward's definition of "sandbox," or in his explanation of self-sustaining gameplay: narrative. There is in fact a lot of narrative that happens in EVE, but it's primarily player-driven. Woodward is skeptical that game-driven narrative can really work successfully in an MMO.

"It seems to me, from my very EVE-centric perspective, that there's a lot of value in narrative gameplay, and there's a lot of value in MMOs," he says. "It's not crystal clear to me how merging those two things together is the best possible way to enhance them both. It seems like they're pushing in different directions, and trying to be different kinds of games.

"For me, and for the way I look at games, those two types of games are not obvious bedfellows, and I wonder whether what kind of compromises you're making when you're trying to get those to mesh together nicely are worth what you're trying to achieve.

"If you're trying to achieve a single-player experience, it's not obvious to me how that's benefiting from a creative point of view from being embedded in a very highly-social multiplayer game."

Related Jobs

Digital Extremes
Digital Extremes — LONDON, Ontario, Canada

University of Central Florida, School of Visual Arts and Design
University of Central Florida, School of Visual Arts and Design — Orlando, Florida, United States

Assistant Professor in Digital Media (Game Design)
The College of New Jersey
The College of New Jersey — Ewing, New Jersey, United States

Assistant Professor - Interactive Multi Media - Tenure Track
Bohemia Interactive Simulations
Bohemia Interactive Simulations — Prague, Czech Republic

Game Designer


Kevin Fishburne
profile image
Agree completely. However, I think MMOs could be "seeded" with initial ideas that the players could sort through and change. The scenario could be initially seeded as an anarchistic wilderness, for example, and players could transform it into a region governed by rules and players enforcing them. If a player vies for power or solace, their goals should be facilitated by the basic mechanics of the game and encouraged as much as any other player whim.

I think the best of scripted stories can't compare to the ones we experience ourselves. "Player-created" stories aren't meant to be Biblical; they just are. Human nature describes itself daily, and so it should in MMO gameplay.

David Navarro
profile image
Brion: Most linear games also lack an interesting narrative. At least with sandbox games you know not to get your hopes up in that respect.

Michael Joseph
profile image

i think the degree to which metagaming becomes a key factor in player enjoyment by adding layers of depth to the gameplay and in playing the game successfully is also important in sandbox games like EVE. This probably falls under the "emergent" category Woodward mentioned.

Openness I think then must refer to how many processes and systems in the game are being simulated at a low level\fine grain as opposed to mostly faked. Minecraft is a great example of how the existance of certain low level simulations allows the player to do unexpected whacky things like construct functioning virtual processors in the game. Gary's Mod is another interesting example.

In sandbox games players get to star in their own episodes of "Have Imagination - Will Travel."

Jakob Gamertsfelder
profile image
Game design seems the sandbox, not so much the games.

Perhaps discussions like this are also part of the sandbox.

Maria Jayne
profile image
I've always defined the term "sandbox" to literally translate into a child in the sandbox. They have a bucket and spade, a set of boundaries that the sand exists and the laws of the world within which to operate. Now using those laws and boundaries the child can make believe whatever they want and construct an extensive although finite reality for themselves.

Now the bucket and spade is a metaphor for any tools given to the user within the boundaries of the world and must abide by it's laws. However the problem arises when you cross your world with someone else s. For example on a minecraft server when you enter and they are building a castle, you might want to destroy that castle, who then has the right? the owner of the server? what if neither player owns the server?

So the owner of the server starts applying additional rules to stop griefing or disruptive behavior, then you add loose objectives for those that choose not to create their own goals, suddenly what was a sandbox is starting to look like more of a set of lego, where you can't do whatever you want, you can only do whatever the pieces you have allow.

Minecraft, to me, was a sandbox, right up until they applied server restrictions and added an end game objective. I'm not sure what I would classify it as now.

I think true sandbox design can only exist within the realm of like minded people, but in order to do that, you have to make rules and restrictions which hinder the nature of what a sandbox is.

Nick Harris
profile image
Cooperatively exploring and colonising a galaxy only to later fight against computer controlled aliens (whether that be carnivorous plants on the surface of a hostile planet, insane pirates, or a full scale invasion fleet), may be more fun. I always have misgivings about killing another player and looting their hard won remains, especially when they have spent months building up their possessions. This is probably why I don't play EVE.

Kevin Fishburne
profile image
A sandbox in its purest form is allowing people to do in-game what they could do in real life if they chose to. Simulating this with a gamepad isn't impossible, but most games take great liberties with player input versus game logic output. I think of it as the ultimate simulation, whether a racer like GT5 or some MMO. User input versus end result is always the premier question.

Bart Stewart
profile image
Emergent, yes; social, if it's a multiplayer game, OK; but goal-directed? That's debatable.

Players will define their own goals in a sandbox game, but the *developer* dictating goals (as in most MMOGs) is pretty much the opposite of the sandbox as usually understood. It's like giving a kid a toy shovel and telling him that he can only use it to dig 5 eight-inch holes in specific locations. Offering an in-game benefit for achieving those goals (or monetizing them) doesn't make that more sandboxy.

Maybe that's the real question: is it misleading to describe a game as a sandbox game if its _primary_ gameplay is designed around developer-imposed narrative or goals?

Chad Wagner
profile image
Can a game offer goals, without restriction?

Give a kid a toy shovel and tell him there is an amazing toy buried out there...and whover has the highest tower in 10 minutes will be rewarded. Do whatever you want.

Is this a sandbox with goals, or is it defacto restriction because you will only be rewarded for select things (unless you are extremely strong willed and are satisfied with your own personal reward system)

Bart Stewart
profile image
That's exactly the kind of thing I was after, Chad -- thanks.

On its face your example seems pretty sandboxy. You're not constraining how to build the highest tower in ten minutes.

On the other hand, what about all the things the player might be doing instead of building a tower? What if he had years to play instead of just ten minutes?

I'm not trying to pick holes in your specific example; I'm interested in the larger point that developer-specified goals tend to reduce sandboxiness. There's no bright line, but eventually you have to say a game that limits player actions to specific goals is so not-sandboxy that it's not longer accurate to say it is one.

There's plenty of gray area. Modern Bethesda games are pretty good examples: Oblivion and Skyrim both create a large, open world filled with "stuff," drop you into it, and say, "have fun!" At that point you can (and I usually do) go wherever and start creating my own stories through my choices of actions. But these games also offer a specific main quest line combining both developer narrative and developer activities. So is that sandboxy or not?

I'd say it's not, not really, not in the way that something like Minecraft has become the ur-sandbox game. The presence of the main questline means that that's how some players will experience the game; for them it's not a sandbox at all. More specifically, open-world games are not about creative, constructive gameplay, which I'd say is a defining characteristic of a sandbox game. You can build your own narrative history, which might fairly be considered a kind of sandboxiness... but that's not what people think of when deciding whether some game is or isn't a sandbox game.

On balance, I think the presence of goals doesn't disqualify a game from being a sandbox game. But that's why I was careful to emphasize the word "_primary_" in my question. If the core of the game is about dictating specific outcomes -- even if the specific steps taken to get there can be player-determined -- then I think I'd say that game doesn't qualify as being called a sandbox game. It might have some sandbox features, but it extinguishes too many possibilities of play to really merit that categorization.

I'm open to other good arguments, though.

Aleksander Adamkiewicz
profile image
I think the distinction is between restrictions, goals and incentives.

You have to decide early on if you want to make a game, or a simulation.
Games like Minecraft are simulations, they are not in of themselves games (speaking in the traditional sense of the word).
The gameplay comes from the players and what they decide to do. I.e. giving the kid a shovel and a (physical) sandbox. It can also be compared to doll-houses and other simulacrums given to children to emulate real-life counterparts.

I think that it doesn't matter if the sandbox features restrictions or incentives as long as the gameplay comes from the player and is emergent from the tools provided (for it to be called a sandbox).

Even if you set up a restriction that you can only use the shovel 10 times, the "gameplay" will be largely dependent on the imagination of the player.

For it to not be a sandbox, the gameplay must be provided by the developer. i.e. you create a game around the sandbox and shovel with rules and goals.
For example you not only restrict that you can only build sandcastles, but that also that the largest sandcastle "wins".

I'm struggling here to describe accurately what I mean because our language is still a bit murky in these matters. Especially the word "game" encompasses far too many contradictory definitions that it becomes really hard to talk about its subgenres (which are equally ill defined).

I would really wished we would draw a clear line between simulations and games.

Per Micael Nyberg
profile image
While I like narrative driven experiences as much as the next guy I would argue that the games are a more snug fit to 'Sandbox' by nature. The iterative creation process lends it self well to an iterative and open-ended product. User to user interaction and user generated content makes for gameplay and business models that scales extremely well over vast amounts of time (at least compared to most AAA console games heavy on narrative). Project Entropia, Club Penguin, TrackMania, Team Fortress 2, Ameba Pico, Second Life, StarDoll etc says hi.

Simon Ludgate
profile image
Comparing a game like EVE to a kid in a sandbox isn't quite accurate. As Bart points out: there are system-imposed restrictions.

Perhaps a better example is a kid sitting next to a sandbox with a remote control car in the sandbox, with one button that, when pushed, operates a shovel on the car to dig out a very specific amount of sand and another button places that sand in a very specific conical shape.

"But I want to build a sand castle with square towers!"

"Sorry Johnny, we didn't design the car to do that."

Ultimately, the "freedom" in the sandbox relies on the tools provided to the players in that sandbox. It's not so much that the developer is taking someone in a free situation and imposing restrictions so much as providing opportunities. It's not binding the normal kids arms, it's giving a rather crude mechanical arm to the amputee.

The problem we face with most "sandbox" games now is a lack of tools. This will be fun, that will be fun, but we can't do anything else because it breaks the game balance.

Daniel Campbell
profile image
CCP really are the masters of emergent gameplay. Dust 514, I felt, was an amazing idea and has an amazing potential. The problem is that the gameplay itself wasn't all that well balanced or interesting.

Raymond Grier
profile image
"Basically it's three things: being social, goal-driven and emergent"

A sandbox game can be open and single player.
Goal driven seems to be a contradiction to what he's defining.

Some people are making comments about the restrictions versus none in a real sandbox, that's not a valid argument. A real life sandbox has constraints: box size, gravity, weight/dryness/stickiness of the sand, what toys are available in the sandbox, etc. Some of these are laws of physics and some are artificial circumstances such as "can my parents afford that Tonka truck I want".

Kevin Fishburne
profile image
Maybe the definition should be the difference between a game with "invisible" versus "visible" walls? With invisible walls, the player just can't understand why they can't do what they want to do and finally chalk it up to some arbitrary limitation in the game. With visible walls, they see the reason behind the restriction and can at least gain some peace of mind if it seems logical.