The Sandbox 2

The Philosopher-Geek has already writ­ten a very good post on the mean­ing­ful game, and that might be impor­tant to read before you step into this one, because I’m going to inves­ti­gate two spe­cif­ic traits that he lists in that post. The spe­cif­ic traits I am inter­est­ed in dis­cussing are: 1. The Meaningful Game does not allow the play­er’s choic­es or pos­si­ble actions to derail the game or con­tra­dict its char­ac­ters, and 2. The Meaningful Game does not con­tain side-quests. These two points are most com­mon­ly bun­gled in games that reflect the “sand­box,” “free-roaming,” or “open-world” style of design, and so I’d like to inves­ti­gate the nar­ra­tive chal­lenges that appear when build­ing such a game, and address the appar­ent slid­ing scale between depth and width, or free­dom and con­sis­ten­cy. But I’ll start by defin­ing what I mean by “sand­box” and “open-world,” and address­ing why such mod­els are pop­u­lar.

Absolute Freedom, Mostly

As has been pre­vi­ous­ly noted in var­i­ous posts, one of the dis­tinc­tive qual­i­ties of games (both video and role-playing) is the inter­ac­tiv­i­ty of a game, or the con­trol that the play­er has over a char­ac­ter, coun­try, and/or nar­ra­tive. Games express this free­dom in a vari­ety of ways, but it’s hard to argue that the great­est expres­sion of char­ac­ter free­dom can be found in games that fit the sand­box model.

Now, open-world games are slight­ly dif­fer­ent than sand­box games, but they reflect a sim­i­lar desire and have sim­i­lar aims. Open-world games are an answer to the “lev­els” of early gam­ing, which offered lin­ear envi­ron­ments expe­ri­enced in a pre-determined sequence. These games more close­ly resem­bled the style of art and enter­tain­ment that came before (mim­ic­k­ing nov­els or films), since the only real free­dom that play­ers could exert was often through the phys­i­cal behav­ior of their avatar char­ac­ter, and poten­tial ends were either “Success,” if the play­er best­ed the chal­lenges of the game, or “Failure,” if the play­er did not per­form ade­quate­ly. Even games that did not have a defined order to lev­els, such as Mega Man, placed sim­i­lar struc­ture before the char­ac­ter. An open-world game seeks to break down the arti­fi­cial bar­ri­ers between dis­crete ele­ments of a game’s envi­ron­ment, instead includ­ing it all in one “world.” However, it is impor­tant to note that there are always lim­its to a game world; all of it must be pro­grammed, and there will always be walls enclos­ing the play­er, whether they are invis­i­ble, illu­so­ry, a level-wrap (think Pac-Man), or enforced by a char­ac­ter’s refusal to leave the area of nar­ra­tive impor­tance. So open-world games are defined more explic­it­ly by their refusal to draw bound­aries between sec­tions of the game and the abil­i­ty to freely explore those sec­tions.

Sandbox games are often syn­ony­mous with open-world games because they essen­tial­ly do to nar­ra­tive con­straints what open-world games do to envi­ron­ments, and it’s eas­i­er for that nar­ra­tive free­dom to be expressed in an open-world than a level-based game. Sandbox games will occa­sion­al­ly do away with the notion of a main plot, but for the most part the plot is avail­able as one option among many, and some­times the play­er can ignore the main plot (and thus the whole notion of “com­plet­ing” the game) to instead pur­sue other tasks in the game-world. Thus, sand­box games often fea­ture pro­lif­ic and prodi­gious side-quests, if they are not com­posed entire­ly of small mis­sions that avoid a main plot com­plete­ly. Sandbox games encounter the same inter­nal para­dox as open-world games, name­ly that there is a limit to what the game can offer. Just as there must be arti­fi­cial con­straints regard­ing the size of the world, there remain arti­fi­cial con­straints in the nar­ra­tive (or non-narrative) struc­ture of sand­box games, such that even­tu­al­ly the play­er will sim­ply run out of things to do or run into sit­u­a­tions where the world can­not be inter­act­ed with in a spe­cif­ic way. So just as open-world games are defined by their refusal to draw bound­aries between “lev­els,” sand­box games are defined by their refusal to draw bound­aries between “right task” or “plot task” and “other game actions.” In other ways, it does not pro­scribe how the play­er should play the game beyond the nat­ur­al con­straints that pro­gram­ming only a cer­tain num­ber of ways that the play­er’s avatar can inter­act with the world gen­er­ates.

This style of game is so pop­u­lar pre­cise­ly because it gives a great deal of a cer­tain type of free­dom to the play­er. The play­er may tack­le tasks that the play­er is most invest­ed in, and in the order that the play­er wish­es. It also allows the play­er to set his or her own pace in the game, which could be a good or a bad thing; I’ll address this in a bit. Essentially, these games offer thou­sands of nuggets of expe­ri­ence, and the play­er is allowed to pick and choose which nuggets he or she will “con­sume”, and in which order. The game is a gate­way to a large buf­fet of poten­tial expe­ri­ences. However, I think that this pro­claimed free­dom can be a trap for the play­er as much as a boon, and I also think it’s worth eval­u­at­ing whether the pur­suit of more free­dom in-game is a wor­thy task, and espe­cial­ly whether this pur­suit gen­er­ates bet­ter art, or at least allows the play­er to either bet­ter access good expe­ri­ences in a game, or access bet­ter expe­ri­ences with­in a game. I’d also like to inves­ti­gate whether this model of game bet­ter lends itself to cer­tain sorts of nar­ra­tive; after all, struc­ture and mechan­ics may ben­e­fit one type of story, and ren­der anoth­er less enjoy­able or com­plete­ly inac­ces­si­ble. By the end of the arti­cle, I hope to iden­ti­fy what sorts of sto­ries the sand­box style best serve, and what types of sto­ries sand­box games should avoid.

The Cost of Freedom

In Polishing the Diamond, Enlightening the Mind, Jae Woong Kim writes “Absolute free­dom is lone­li­ness.” I think that this describes my aver­age expe­ri­ence with sand­box games. Take Oblivion as our first exam­ple. In order to give the play­er max­i­mum free­dom in the pro­duc­tion of this char­ac­ter, the play­er starts the game by gen­er­at­ing a char­ac­ter of any race and with any set of favored abil­i­ties (or ways to inter­act with the game), regard­less of whether these choic­es are intel­li­gent deci­sions or not. The choice of race does have some small influ­ence on your abil­i­ties (some­times a big influ­ence), but it has very, very small nar­ra­tive reper­cus­sions. Your abil­i­ties can deter­mine what nar­ra­tives you have access to in-game, but the choice itself has no impact on the nar­ra­tive.

In almost all ways, your char­ac­ter is a blank slate; more­over, the tools that the game give you to flesh out that char­ac­ter’s per­son­al­i­ty are not very diverse, nor does the game as a whole take much notice. In the pur­suit of giv­ing the play­er absolute free­dom in his or her envi­ron­ment, the depth of dis­tinct nar­ra­tives (that con­tain spe­cif­ic main char­ac­ters, a uni­fied theme, and a cohe­sive “plot”) tends to be shal­low, short, and emo­tion­al­ly stunt­ed. It is hard to lay this fail­ure at the feet of the writ­ers; after all, they’re respon­si­ble for thou­sands of non-player char­ac­ters. In an attempt to give as wide an expe­ri­ence as pos­si­ble, and to pro­vide such a diverse range of expe­ri­ences (includ­ing rad­i­cal­ly diver­gent sys­tems of mechan­ics with­in a sin­gle game, such as a stealth sys­tem, magic sys­tem, melee com­bat sys­tem, ranged com­bat sys­tem, lev­eled sys­tems for every­thing from alche­my to armor, and the attempt­ed uni­fi­ca­tion of these sys­tems) and nar­ra­tive sup­port for each expe­ri­ence means that the game is spread thin. If more focus is given to any one por­tion of the game, then whole sys­tems might have got­ten the axe… prob­a­bly includ­ing the “Axe” skill.

The play­er never has a dearth of options in Oblivion, but unfor­tu­nate­ly every option pro­vides a nar­ra­tive about as ful­fill­ing as any other game’s side-quest. Oblivion falls short of being a Meaningful Game sim­ply because the play­er rarely, if ever, feels that any­thing mean­ing­ful is going on, and his play­er char­ac­ter is mean­ing­ful and inter­est­ing only because the play­er inhab­its him or her. This would be less of a prob­lem if inter­est­ing char­ac­ters pop­u­lat­ed the world, but they don’t.

If you’re unfa­mil­iar with Oblivion and games like it, here is an exam­ple of a sin­gle quest. The game is absolute­ly full of sit­u­a­tions like these, and it dis­plays the level of depth that all but a few char­ac­ters in the game dis­play. I am fair­ly cer­tain that the design­ers of Oblivion were more inter­est­ed in cre­at­ing the broad expe­ri­ence of “There is a whole world out there wait­ing for me to inter­act with it in minor ways” than the expe­ri­ence of “I am a per­son doing mean­ing­ful things.” This is gen­er­al­ly true of most sand­box games; the veneer of free­dom and option comes at the sac­ri­fice of depth, emo­tion­al­i­ty and mean­ing.

Oblivion lends cre­dence to the depth vs. width scale, that if a game grows wider in options and con­tent, then that con­tent will be less deep and reac­tive. While this scale isn’t a given, it is use­ful in express­ing why the events of the game are not often com­pelling or engag­ing, at least when assessed as a nar­ra­tive. While play­er free­dom is what makes video games a unique art form, if that free­dom is viewed as an end (high free­dom equals a good game) instead of a means (play­er choice is a vehi­cle through which we can tell a com­pelling story suit­ed to the medi­um), then it will not pro­duce a Meaningful Game.

The fact is that it is dif­fi­cult to cre­ate a video game with a com­pelling nar­ra­tive or, per­haps more accu­rate­ly, few video games suc­ceed at pre­sent­ing a com­pelling nar­ra­tive. I also believe that, for a game to be mean­ing­ful (and qual­i­ty art), its ele­ments must be qual­i­ty. If such a game has a nar­ra­tive, then it must be a qual­i­ty nar­ra­tive. It’s my opin­ion, then, that if a sand­box game presents a sin­gle, “main” nar­ra­tive, then the entire sand­box must exist with­in the nar­ra­tive in order for the game to be mean­ing­ful, and prob­a­bly in order for the game to be qual­i­ty art. There are also alter­na­tives to a sin­gle main nar­ra­tive, which I’ll get to in the next sec­tion.

Oblivion is indica­tive of most sand­box expe­ri­ences in that, while it is the prod­uct of a great deal of work and effort, and is cer­tain­ly admirable in many respects, it fails to be mean­ing­ful due to its shal­low­ness and the deriv­a­tive qual­i­ty of its con­tent. In other words, for all their scope, and arguably because of their scope, sand­box games tend to be bad art.

Paper Beats Rock, Sandbox Beats Story

The sand­box is pop­u­lar for a rea­son: there are a lot of things to like about the sand­box. Player free­dom IS inter­est­ing, and the abil­i­ty to set your own pace and access the con­tent that you find inter­est­ing soon­er rather than later can be a good thing. Moreover, I don’t think that the nar­ra­tive fail­ings of most sand­box­es are a trait of the sand­box, but it does seem to indi­cate that simul­ta­ne­ous­ly cre­at­ing a cen­tral nar­ra­tive and a sand­box is a self-defeating route. The typ­i­cal nar­ra­tive for­mu­la seen in most games clash­es with the notion of absolute play­er free­dom, either gen­er­at­ing mas­sive incon­sis­ten­cies in char­ac­ter or plot, or sim­ply forc­ing the char­ac­ter or plot to be so empty that they mere­ly serve as a rough moti­va­tor for the events of the game (or the plot that’s there because they said we need­ed a plot), which both gen­er­ate bad art.

In some tales, the sand­box style sim­ply does not make sense. In any nar­ra­tive in which there is an impor­tant cen­tral nar­ra­tive and the side-events detract from pur­su­ing a larg­er, more press­ing issue, a sand­box does not suit the nar­ra­tive. The first arti­cle on the mean­ing­ful game spoke on side-quests, and the points that Bill made regard­ing them holds just as true for the way that sand­box­es inter­act with cen­tral nar­ra­tives, so I won’t tread over that ground again. For the most part, then, I will argue that sand­box games should not attempt to fol­low any sort of lin­ear nar­ra­tive; they sim­ply aren’t suit­ed to that type of story. Infamous, Fallout 3, Grand Theft Auto… these games are all fun, but they all have a cen­tral nar­ra­tive that falls flat because of its loca­tion in a sand­box, and because the devel­op­ers decid­ed that hav­ing a sand­box was more impor­tant than mak­ing the story worth­while and com­pelling.

Assassin’s Creed (the first one; the sec­ond begins to lose this trait) is an exam­ple of a game that suc­cess­ful­ly builds a cen­tral nar­ra­tive into a sand­box style game. This is because the nar­ra­tive is in no way sac­ri­ficed to play­er free­dom, because Altair’s goals are always para­mount and are the source of all his actions, because even “filler” activ­i­ties often use the oppor­tu­ni­ty to fur­ther char­ac­ter­ize Altair and the world he lives in, and because there is very lit­tle in the game that isn’t designed to belong with­in the “cen­tral” expe­ri­ence and nar­ra­tive. Hunting down flags and ran­dom tem­plar begins to feel a lit­tle game-like, but the game pre­dom­i­nant­ly rewards actions that drive the game along, thus main­tain­ing a prop­er pace, and, more impor­tant­ly, never inter­rupt­ing the dis­tance between the play­er and Altair. Most sand­box games do not fea­ture so stel­lar an imple­men­ta­tion; Assassin’s Creed’s focus on story and con­sis­ten­cy is the excep­tion.

Therefore, I argue that in order for the sand­box to be mean­ing­ful, the devel­op­ers must avoid sim­ply slap­ping a typ­i­cal, straight­for­ward, novel- or blockbuster-style nar­ra­tive onto a sand­box and expect­ing it to be suc­cess­ful. Different forms often demand a change in con­ven­tion and struc­ture, and the sand­box rarely receives prop­er treat­ment in this regard; instead, if a nar­ra­tive exists, it is a typ­i­cal, straight­for­ward affair strung like a thin thread through the mid­dle of the sand­box, and its qual­i­ty is usu­al­ly poor or, at best, rough­ly equal to the qual­i­ty of the nar­ra­tive in the rest of the game.

But It Doesn’t Have To Be This Way

Sandbox games could become much more mean­ing­ful if they fol­lowed Assassin’s Creed’s exam­ple, but the truth is that the tra­di­tion­al nar­ra­tive model does­n’t func­tion as well in a sand­box (even AC has reg­u­lar slip-ups, becom­ing deriv­a­tive here and there). Luckily, there are many kinds of story, and many ways to incor­po­rate story into a sand­box.

Right now, the most obvi­ous model of hero­ic story-telling is what we see in nov­els and action/adventure films. These are long-form sto­ries with high stakes, and they typ­i­cal­ly fea­ture a con­sis­tent build in ten­sion and con­flict until the “final con­fronta­tion,” in which the hero van­quish­es the vil­lain that has tor­ment­ed him or her for so long. It is this model, the sin­gle, unin­ter­rupt­ed, build­ing hero nar­ra­tive, that is most com­mon­ly seen in games, and sand­box games often wind up with this sort of nar­ra­tive sim­ply because games today need nar­ra­tives, and stick­ing to the mold is either the obvi­ous choice, the only per­ceived option, or the choice that seems like­ly to sell best, since it has worked so well in near­ly every other game.

But if we step out­side of video games for a moment, we can see that there are many ways to tell a hero nar­ra­tive; heroes don’t just exist in nov­els and feature-length films. The mod­els exhib­it­ed by tele­vi­sion shows (i.e Buffy, Star Trek, Veronica Mars) and col­lec­tions of short sto­ries (i.e Elric and Conan) offer a pow­er­ful alter­na­tive to the feature-length nar­ra­tive, and even con­tain meth­ods of break­ing up a cen­tral nar­ra­tive or doing away with the cen­tral nar­ra­tive alto­geth­er. Conan sto­ries are strung togeth­er by the char­ac­ter, noth­ing more; Star Trek: The Next Generation is a con­tin­u­ous tale because it deals with the same char­ac­ters, but for the most part each episode is a reset. There is some sense of “whole story,” in that the indi­vid­ual tales are linked, but the tales can be accessed indi­vid­u­al­ly, and there is no design for a grand nar­ra­tive that builds to a huge pay-off.

In my opin­ion, the tropes and mod­els of these short-form sto­ries pro­vide a much bet­ter rubric for sand­box games than do cen­tral nar­ra­tives. Instead of declar­ing one par­tic­u­lar series of events the “main” story and mak­ing every­thing else filler, this model would real­ize that a sand­box game is real­ly try­ing to tell a whole host of sto­ries, but would remove the temp­ta­tion to let the nar­ra­tive’s qual­i­ty slide. I expect that, in order for the aver­age sand­box game to be a mean­ing­ful game, the devel­op­ers will have to take cues from tele­vi­sion shows and short sto­ries, and so seek to tell tales of lim­it­ed scope and, hope­ful­ly, with a con­sis­tent­ly high value of emo­tion­al con­tent and mean­ing.

I’m curi­ous to know what you think about this idea. I sus­pect that a par­a­digm shift is what’s required for sand­box games to become mean­ing­ful and man­age­able; do you agree? Or do you think that the prob­lem lies else­where?

Matthew Schanuel

About Matthew Schanuel

Matthew Schanuel lives in Boston, Mass. He's a beer aficionado, a game player (and designer!), an academic-in-exile, a DM, and, most recently, an employee of a financial non-profit. He draws the comic Embers at night over at

2 thoughts on “The Sandbox

  • TJ

    I think all of the points you made are great. I would say that attach­ing a nar­ra­tive in the form of short sto­ries or episodes is the best pos­si­ble path. It seems as though GTA actu­al­ly attempts to do this, but most­ly fails. I’m not sure about Assassin’s Creed though because when I played it it felt very gamey. I could­n’t help but think each flag col­lec­tion mis­sion was a chore. I should revis­it it at some point.

    With that aside I’m curi­ous as to why sand­box play­ers actu­al­ly enjoy the games so much. You men­tion this briefly, but I won­der if cur­rent play­ers are actu­al­ly the prob­lem. Current play­ers might reject a mean­ing­ful game. Obsidian play­ers, for exam­ple, have found ways to play the games for hun­dreds of hours yet I have never been able to sit down and play it for more than an hour.

    Instead of enjoy­ing nar­ra­tive, I think they take a com­plete­ly dif­fer­ent path. One which might be pure sim­u­lacra (am I using that right?), that is, enjoy­ing a game pure­ly for it’s ele­ments of game and fan­ta­sy. They enjoy max­ing out their char­ac­ters, gain­ing new equip­ment, see­ing orcs, col­lect­ing runes, look­ing at the cool graph­ics, etc. They enjoy doing all of this with no story (ala MMORPG). In fact, a TOTAL lack of nar­ra­tive might even be prefer­able to them (see Minecraft). If there is no nar­ra­tive, there is noth­ing to remind them that what they are doing is a point­less grind with­in a sys­tem. Similarly, there is noth­ing to remind them of grander nar­ra­tives (stop play­ing so much Obsidian and get a job, etc.). This is fur­ther dis­played by the types of mods that are pop­u­lar with seri­ous play­ers. The pop­u­lar mods never add nar­ra­tive.

  • Matthew Schanuel

    I def­i­nite­ly think that sand­box­es do cer­tain things well, and they can be very fun, but they’re not very “mean­ing­ful.”

    I know that you’re cor­rect in that gamers play games for a vari­ety of rea­sons, and I don’t expect the sand­box to become any less pro­lif­ic, nor do I think it should, nec­es­sar­i­ly. But if one want­ed to make a sand­box game a mean­ing­ful expe­ri­ence, one would make it like this. I also antic­i­pate that such a game would prob­a­bly be pret­ty well received; qual­i­ty games some­times get the praise and sales they deserve.

Comments are closed.