[The second in Gamasutra's 'Game Design Essentials' series, following '20 Difficult Games', looks at the roots and design lessons of 'open world games' - titles in which the player "is left to his own devices to explore a large world" - from Adventure through Metroid to Grand Theft Auto.]
Defining 'Open World Games'
When we discuss "open world games" in this article, or sometimes "exploration games," we mean those games where generally the player is left to his own devices to explore a large world. What all of these games share is the seeking of new, interesting regions at whatever time the player deems fit. No force forces the player's motion into new areas. There's no auto-scroll, and there are no artificial level barriers.
A couple of games, such as Air Fortress, push this by offering many mazes, but they are more like many individual games than one large map. One of these games, Cadash, is presented for contrast, but technically fills the description. Most of the games take closely after Metroid games, or are one of them. Fully half of the games here contains substantial side-view platformer elements.
At the core of the open world game is consumption. Once a place is seen for the first time, it cannot be unseen and seen again. To an extent, the game is a maze, and once the whole thing is seen the game cannot be played the same way again. Perhaps it can be played for a good score or a good time, but that's a substantially different kind of experience.
Some games attempt to offer replayability through randomization. Roguelikes and strategy games in the mold of Civilization do this. Some allow the player's state to vary, allowing them to reach a given point in the game world in a variety of ways, each with its own implications for the situation found there. However, no game in this list uses randomization in this manner.
Notes On The Chosen Games
Some particular comments on the games that were picked for this list:
1. As with all such lists, some things had to be left out. Morrowind and Oblivion are particular games in which exploration plays a prominent role, and their absence, or that of any other game, should not be taken as a slight against them. There are other reasons they are not mentioned here, one of them being that I think it's best to possibly save them for a later article on RPGs.
Also notably absent are any adventure games, either textual or graphic. This is also not intended to short that genre. Also missing is the notable Commodore 64 game Phantoms of the Asteroid, but it was covered last time. Another Commodore game that could make the list is Spindizzy, but the fact is I don't think I have enough experience with that game to write well about it.
2. As with the previous article, 'Game Design Essentials: 20 Difficult Games', this is not intended to represent the best open world games, or the ones that are most "explore-y," although some of them are pretty nice. The games are chosen for their instructive qualities and general interest, not to compare them using a meaningless yardstick. They're here because I could illustrate something important using them as examples.
3. Some out there in Internet Land snarked, concerning the previous list, that it was biased towards older games. Yes it was, and I make no apologies. Older games tend to have more elemental designs, presenting their mechanics strongly rather than submersing them between a sea of what a game is "supposed to be." This is particularly useful for explaining and highlighting design conventions.
1. Adventure (2600)
The first action-adventure game.
Published by Atari.
Designed and developed by Warren Robinett
Platform: Atari 2600
From a design standpoint, watch for the fact that, on a system with 128 bytes of RAM and completely on-the-fly, raster-based graphics, each of the objects in Adventure is still more interesting than anything contained within 99% of current games. It is the most scathing indictment of "modern" gaming possible.
Further: Everyone knows, by now, about the famous easter egg. In case you don't: on higher game variations, there is a one-pixel-size dot in a locked-off room that's the same color as the background. Use the bridge to get into that room in order to pick up the dot, then take it to the room that has a black line on its right-hand edge and drop it. Bring enough other objects to that room and the line will flicker. At that time, you can move through the wall by pressing against it to see a graphic of Warren Robinett's name. Consider what percentage of the game's ROM is taken up by that graphic!
That anything other than Pong could be made on the Atari 2600 is rather incredible. By default, the system had four kilobytes of ROM space and 128 bytes of memory. Its graphics capability was more primitive still.
Atari's developers, at the time of the system's heyday, amused themselves with trying to push the system farther and farther. The creator of Adventure also wrote a BASIC interpreter for it. It was just barely useable, but it worked. Someone else wrote a chess player for it. The system is so challenging to write for that in recent years 2600 programming has become something of an extreme sport for geeks.
The 2600 was designed to play Pong-like games, but as is often the case, amazing things are possible if no one tells you what can't be done. So it was that Warren Robinett, despite having been told by his bosses that the project was too difficult to even attempt, decided to condense Crowther and Woods' classic text adventure, Adventure, to the system.
He didn't succeed in porting it, or even really capturing its flavor, but that's okay because the resulting game is probably the Atari 2600 game that stands up the best today. Even Pitfall II isn't quite as fun to play today as it was when first developed, but Adventure is still quite entertaining to run through for a few minutes.
The player is represented by a square, and his abilities can be summed up as being able to:
Pick things up by running into them.
Drop things by pressing The Button.
Return to the start location (by pressing reset, useful if eaten).
Get eaten by dragons. (Actually, the player doesn't have to do anything for this to happen. It more or less occurs on its own.)
Yet look at all that can happen:
Castles can be unlocked, by touching their doors with the right key.
He can also re-lock doors the same way.
Mazes can be explored, by moving around them.
Dark corridors can be explored because the area around the player lights up.
Dragons can be killed by touching one with the sword.
Some dragons are also afraid of things. The yellow one is afraid of the yellow key, and it can be used to chase him off.
The bridge item allows the player to pass through any horizontal wall it covers, and is used by dropping it in the right place.
The bat steals items, carried or not.
The bat can also move dragons around, sometimes producing moments of sudden peril.
Since the bat behaves randomly, he might also steal away a dragon that's chasing the player.
The bat might even bring a useful object to him.
The player, tiring of the bat's antics, can pick up the bat himself and carry him into a castle. If the bat or a dragon is inside a castle when its door is locked, it's trapped there!
He can use the magnet to pull objects through walls. Again, the magnet works entirely by proximity.
If the player gets eaten, and the dragon gets picked up by the bat, the player gets carried along, treated to an aerial view of the game world. (It's not too useful, but is interesting.)
Finally, the player can win the game by bringing the chalice to the Yellow Castle.
Making the game's triumph complete is the random adventure mode that scrambles the locations of the objects at the beginning of play. It's been noted that one game in eighteen is unwinnable in this mode, but considering that it's on a freaking Atari 2600, I think that kind of fault can be excused for once.
For such a small game world (there are only about a couple dozen screens in all) the game shows remarkable ingenuity in making the most of that space. Initial exploration of the world doesn't take long, but it's the varied terrain combined with the effects of the objects (including monsters) that make it interesting.
Adventure's fun comes from the way all of its simple objects interact to produce complex behavior. Carried objects continue to operate, whether it's the bat or player who holds them, so the bat might carry the magnet through a room where the player is using the bridge, moving it out of position and forcing him to find another way back. Or, carrying the sword, the bat might brush it across a dragon on his flight, killing it. This is possible because all of the objects in the game function automatically, which they have to be anyway since The Button is devoted to dropping stuff. A lot of the fun in Adventure comes from the unintended consequences of the player's actions.
Further, while the game contains what can only be described as a tiny game world, it is considerably replayable. Game variation 3 scrambles the locations of the game objects and monsters at the start of play, and the various implications of those objects can sometimes produce special challenges, like the sword being locked inside a castle, or the dragons appearing clustered together.
Indenture, a freeware DOS recreation of the original game with some extra features
An action-adventure that works its isometric perspective unusually hard.
Published by Sega
Developed by Climax Entertainment (Japan)
Platform: Genesis, Wii (Virtual Console)
Landstalker is a game that revels in its isometric presentation. Other isometric jumping games (there were more than a few from Europe) worked to keep each screen easy to understand, so the player wouldn't get confused as to how the platforms of each area related spatially to each other. There are places in Landstalker, on the other hand, that seem as if they were created specifically to be optical illusions.
While it is indeed its own game, at first glance it looks like an isometric knockoff of Zelda. Your character is an elf, is dressed in green, hangs out with fairies, explores a lush landscape, and frequently collects heart containers, here called "life stocks."
Where it diverges from the mold is in its vibrant characterization. Every character in Landstalker has a distinct, often humorous, personality, including the protagonist Nigel. The characters are memorable to a degree little seen outside the Grandia games, or perhaps the original Shining Force, also developed by Climax. It is truly a game that contains no angst. Nigel and pint-sized assistant Friday aren't even in it to save the world; their aim is to get filthy rich, and while they do good along the way it is fitting that the conclusion of the game results in them getting showered in gold coins. They may be treasure hunters, but they work for their loot.
And oh, how they work! Despite its Japanese production, Landstalker is really a descendent of those European isometric jumping games, dating back to Airball. Every one of them is maddeningly difficult, and not just for reasons dealing with the perspective. Other games of the type include Spindizzy and little-known SNES sequel Spindizzy Worlds, Head Over Heels, Light Crusader (a very atypical game from Treasure), Taito's arcade RPG Dungeon Magic (a.k.a. Light Bringer) and Sony Imagesoft's Solstice and Equinox. Landstalker is nothing less than an isometric platformer, infuriating puzzles intact, expanded and made into the basis of an entire game world.
It turned out pretty well. The joy and humor with which the characters are written and presented serves as a nice counterpoint to the difficult jumping challenges. While the game is one of the more linear examples on this list, it doesn't push the player to make progress. In Zelda style, there are "Life Stocks" hidden everywhere which serve to increase Nigel's life bar, and players will find themselves well-rewarded for poking around. And the game sports a very nice variety in setting. One area is a sequence of devious riddles, and another is a gigantic hedge maze. These places are important for exploration-type games where the vistas are much of the point.
Landstalker is one of those games where you hope, at the end, for another game with the same characters, which of course never happened except for Nigel & Friday's inclusion in Time Stalkers (a.k.a. Climax Landers), a suspiciously indulgent pseudo-roguelike in which Climax took characters from a number of other games they'd made and shoveled them all into a generic setting.
Most open world games make the player's character an unknowable cypher, purposely without words in order that the player can project his own thoughts onto him. But Landstalker takes the opposite approach, with effectively two protagonists who each comment during conversations. It is hard to imagine more likeable game protagonists without straying over into Game Arts' territory.
Landstalker also proves that difficulty can make an open world game better. There are traps on Mercator Island that will make any player want to throw down their controller, but the sheer variety of situation, just of seeing what the next area will throw at him, is enough to make him want to continue. The riddle area by itself is intriguing enough to warrant a play-through.
3. Pitfall II: Lost Caverns
Possibly the first big-world game.
Published by Activision
Designed and developed by David Crane
Platform: Atari 2600
Length: Short (Medium if played until mastery)
The Atari 800 version has a second quest!
This is the most advanced game on the old 2600. It's certainly so technically: it's got a large mapable, sometimes scrolling world with independently-functioning enemies, special regions, multiple modes of movement (running, jumping, swimming, ballooning). It's even got pretty good background music that changes based on the situation.
But its design is even more advanced than its programming tricks. While the original Pitfall's exploration-based gameplay was tightly limited by its three lives (for novice players) and 20-minute time limit (for advanced players), Pitfall II discards entirely lives, health, and even death. When the player touches a fatal enemy, the music changes for a bit and he's transported back to the last "cross" he touched, and he loses points depending on the distance. Crosses aren't too common, but they aren't too rare either.
The main thing this did for the game was remove the requirement of having to be really good to see it all. The original Pitfall had 255 screens and 32 treasures, of which probably less than 1% of owners had ever seen all of. Pitfall II, while still challenging, could be played continually until finished, with the player's remaining score still providing a substantial measure of skill that could be improved through further play.
The result is, interestingly, not dissimilar to Metroid without weapons or power-ups, but with a score. Pitfall II even has that game's search aspects: instead of looking for missiles and new abilities, players search for treasure. Not bad at all for a 2600 game!
Realizing that game-ending conditions other than winning could be discarded entirely was an amazing insight, and possibly marks the origins of the structure seen in 95% of games to this day: play until death, then send the player back to a checkpoint or previous save. While most of these games will boldly declare "Game Over" when the player runs out of health, considering you can always resume from the last checkpoint it's really not over at all. It's arguably an overused design now, but it wasn't back when Pitfall II did it.
Notice that Pitfall II leans even harder on score than Pitfall! did. While the origin of the concept of score goes back to pinball and earlier, is an oft-neglected aspect of gaming these days. Sure, there are "experience points" and "hit points" and "magic points" and a dozen other points, with more introduced every time a Japanese RPG tries out a goofy new system, but most games have shied away lately from providing measures of skill. Exploration games, in particular, tend to focus more on finishing the quest than building up a score, but Pitfall II shows that they are by no means incompatible.
Not the first open world game, by a long shot, but refined the concept and introduced the Metroid structure that rules action-adventure game design.
Published by Nintendo
Developed by Nintendo R&D 1
Designed by Gunpei Yokoi and Yoshio Sakamoto
Platform: Famicom Disk System, NES, Gamecube (Metroid Prime), GBA (Metroid Zero Mission), Wii (Virtual Console)
Take a look, the game is composed entirely of horizontal and vertical-scrolling areas. They almost always alternate, that is, horizontal areas always have vertical areas connected to them, and vice versa. There are only two types of places where this trend is bucked: powerup rooms always connect horizontal-to-horizontal, and the final boss room is also connected this way.
Also consider for a moment the Maru Mari, a.k.a. Morph Ball, one of the strangest powerups in gaming history. Mario growing large from eating a mushroom is an obvious nod to Alice in Wonderland, but where the hell did this come from? It has become one of the most identifiable aspects of the Metroid series, enough so that the Prime games had to include it. When game designers sit down with a notebook and start jotting down abilities for their kick-ass warriors to have, turning into a small ball on command is not what one expects to see written down. Even less does one expect to see it survive the editing that brainstorm sessions demand. That it did, and that it became one of the signature elements of the games, says much for the elemental power of creativity in game design.
Metroid was not the first open world game, and neither was it the first side-view platformer exploration game, nor was it the first game where players found things in the maze to allow them to reach new sections. But it was likely the first game to take these different elements and rigorously mold them into a game-ruling structure.
Tellingly, the game starts out by immediately forcing players into a wall imposed by its structure. The first room (I'll call a single horizontal-scroll or vertical scroll area a "room") contains a powerup item, the Morph Ball. The third room contains a low ceiling that cannot be passed without it. If the player tries to tackle Metroid like Mario, always going to the right, he'll hit the barrier immediately; the Morph Ball can only be obtained by going left from the start.
In effect, by putting such a barrier in the first two rooms of the game, the designers are telling the player:
1. You can explore in more than one direction.
2. There are cool things in this game that give you permanent new abilities.
3. You'll need them to progress.
4. If you find a place you can't get by, go back and look for new powerups.
Modern games tend to be chatty enough that, were the game made today, all this would probably have been printed on the screen at some point, if not voice-acted. But Metroid is silent. Outside the opening and ending, there are only six words visible in the entire game. (They come near the end....) At the beginning of the game the player's explored area is limited, so there's not a lot the player can do other than go back and find the Maru Mari. Later on the player is more likely to be without a needed power-up, but by putting such a block almost at the start the player comes to learn the rule of powerup progression for himself.
The atmosphere is the thing about Metroid that holds up the best today. I'm not the first to remark that Metroid's awesomeness comes, by large part, from the fact that the game doesn't seem designed. It's easy to look at the chaotic arrangement of tunnels, large unimportant sections, frequent dead-ends and random strangenesses (like the "fake" version of boss Kraid that lurks in one tunnel) as signs that the game was created by computer. It's actually intricately planned out, but exploring the planet makes it seem like the level designer was determined to erase any signs of human handiwork. This makes Metroid what one might call a Lovecraftian game. It is easy to believe it the work of inhuman logic, built by beings unknown to us.
It's dangerous to say definite things about whichever game did something first, but the following seems safe enough. Metroid gave us the first large-scale use of granting permanent player abilities as a means of game progression. It gave us major powerups (Long Beam, Hi-Jump Boots, Screw Attack, etc) as both rewards and an advancement system, while also including many minor powerups (Missiles and Energy Tanks) that serve as simple rewards. Yet they too are essential objects; Energy Tanks of course extend the player's maximum health, but large numbers of Missiles are eventually needed to destroy the Zebetite barriers in one of the last rooms of the game. If the player hasn't found enough missiles he's stuck, lending importance to what might otherwise be a trivial collection goal.
Metroid Cubed, a fan recreation using a voxel engine
One of three games on this list that first saw life in arcades, a market so anathematic to exploration goals that it's notable when any game there features them.
Developed and published by Taito.
Length: Very short.
Many supposed RPGs pay mere lip-service to the conventions, especially if they're action games. In Capcom's two Dungeons & Dragons brawlers, for example, players advance in level automatically as they finish levels, and experience points are merely for score. Taito's two 90s arcade RPGs Cadash and Dungeon Magic (a.k.a. Lightbringer), on the other hand, attempt as best they can to incorporate all the RPG aspects they can into their reflex-testing, real-time worlds. You can even take the time to talk to townsfolk, although the strict timer would seem to make that a dubious proposition at best.
Taito is one of my favorite developers. They don't get the fandom that Nintendo or Sega does, but in their heyday they were just as inventive as any of them. Recent years have seen them fall on hard times, and subject to the ignominy of being bought by Square/Enix, who has just about as little to do with classic arcade games as any game manufacturer.
Cadash is a game that fits my "open world game" definition only with effort. It is true that it is a large-world game, but it is also a fairly traditional action RPG, one made notable in large part for its appearing first in arcades instead of in console ports. (It got two of those.)
Most RPGs don't fit into the exploration game mold because, although generally nothing stops the player from going back to old areas, there is usually so little reason to that it's a waste of time to do so. The basic RPG structure is: fight monsters in an area until the player has strength to beat its boss, beat said boss, then proceed to the next. Each area has stronger opponents than the last, and each provides greater rewards for beating them, so it's always numerically advantageous for the player to bumble around the most advanced area available to him. The monsters may be tougher, but they're worth so much more experience and money than those in previous areas, and the points required to attain the next level are great enough, and the equipment that can be found or purchased in the n