Hi, I'm Steve Gaynor. I am a level designer. I worked on BioShock 2 at 2K Marin, and was
writer & lead designer of the Minerva's Den DLC for that game. I'm at
Irrational Games now, working on BioShock Infinite. Whereas you...
...are standing in an open field west of a white house, with a boarded front door. There is a small mailbox here. You approach the door and a man peeks out from between the boards. He says, “Only members of the Order are allowed in this house. Bring me proof in the form of a Seal of the Order
and I shall let you inside.” Hmm. You know what? Scratch that.
Let's try again.
You are standing in an open field west of a white house, with a boarded front door. There is a small mailbox here. Moving to the side of the house, you find a cellar door. Its keyhole is engraved with a distinctive dragon crest. You open the small mailbox.
Inside you find a key, the head of which is carved with a dragon emblem... Alright, no.
One more time.
You are standing in an open field west of a white house, with a boarded front door. There is a small mailbox here. You head to the back yard of the house and discover a toolshed. Inside the shed, you find a heavy ax,
recently sharpened. You take the ax and return to the front of the house, chopping through the boarded door and making your way inside. Alright, I think that’s a keeper. I’m going with that one. So, why? What in the world do these alternate visions of Zork
have to do with the games I've worked on, and what you might be working on now or in the future? It's a question of values.
The games I'm drawing on for this talk are part of a subgenre that’s often referred to as the “immersive simulation,” which as a genre is defined by a fairly specific set of values that are nevertheless hard to pin down. But I’ll give you my personal interpretation.
For me, these are games in which I seamlessly inhabit an avatar.
In which I mine the stories out of the world.
Where I grow my abilities to more capably thrive.
Where I become fully enmeshed in an imagined place. It's only one specific approach amongst the many out there, and a fairly rare one-- it's broader than your pure action FPS that's super focused on an intense combat experience, but narrower than your full-blown RPG that's about true open worlds and deep statistics.
But that's where the “simulation” idea comes in: I'd say many kind of games can be “immersive,” but the “immersive sim” attempts to place the player in a specific role and place, and give them the ability to do whatever their character might logically be able to do within that context.
Interactive believability is paramount to this approach, which I think is one thing that sets the subgenre apart. But if many different kinds of games can be “immersive,” what is the inherent value of immersion? It's a common buzzword in game development, but there's another more universal term for it:
suspension of disbelief-- that the audience forgets about anything but the fictional world, at least for a little while. We craft incredible places for players to visit-- in shooters, in roleplaying games, in adventure games and more. Our foremost goal should be maintaining the sensation that these places are real, and that the player feels as if they are really there, in the moment of experience. It is a vital illusion, and it is a fragile one. It is, in many ways, a question of naturalism.
We can build worlds that look and feel naturalistic, sensorily speaking. Visuals, audio, effects. But these worlds contain gameplay, and games are defined by consistent sets of rules
-- patterns of behavior. The human mind is very adept at detecting patterns, and so our rules must maintain a naturalism in their systems that supports the naturalism of the space. It's a question of being consistent, and therefore predictable, but not TOO predictable. Any pattern of behavior that is too overtly mechanical will feel artificial; conversely, any special case behavior that defies the rules will immediately stand apart as breaking expectations. When the player's mind detects that we've strayed outside either side of this careful balance, it betray the designer's hand. The carefully constructed gameworld is no longer a real place; it's a false experience made by some guy.
“ Some guy” should not be part of the equation.
We want our players to forget us. I believe that the above balance can be struck in the systems space. But I'm a level designer. As storytellers and experience builders, we construct worlds that need to be experienced in order
we build level progression around a carefully tuned difficulty curve, and present a story with a beginning, middle and end. As far as naturalism goes, this is an aberration. If you were to walk out of this room right now-- not that I'm suggesting you do-- there would be nothing keeping you from walking north, south, east or west out of the building, ending up in Canada or Pennsylvania or the Atlantic. The friction between these elements-- a naturalistic, believably crafted world, governed by consistent rules, twisted into a linear progression-- is often the culprit when suspension of disbelief is broken. The special case conditions that enforce linear progression stand out from the game's rules and make the artificiality of the experience obvious. Before the gatekeeper will let you through the door to the next level, you must present him with the Seal of the Order
The locked door will only open when the Dragon Seal Key
has been retrieved from its conveniently-accessible hiding place. These objects lie outside the rules of the gameworld. They are disempowering, unyielding, within a world where the player is supposed to be in control of their experience. When gating is transparent, the player is not only thinking about us, but the more design-conscious among them probably resenting us.
While there's no magic bullet, we should do everything in our power to reduce the inherent tension between believability and linearity. I'm going to talk about one method of controlling access to game spaces in an authored order that is consistent with game rules and empowers the player. The concept is staged tool gating.
Just to get the definition out there, gating is the practice of restricting access to sections of the gameworld until certain prerequisites are met. It is the basic building block of authored progression. These prerequisites can be practically anything. They are your “objectives,” your “mission goals,” or your “quest steps.” They're the specific conditions that keep you from moving forward along the critical path toward the end of the game. The simplest example might be in The Legend of Zelda:
upon first entering any room in a dungeon, the doors of the room close and lock. Upon killing all of the monsters, the doors reopen, allowing you to continue. Gating mechanisms can be categorized in a few simple ways.
Local vs. Distributed, Systemic vs. Special Case, and Intuitive vs. Unintuitive. This gating mechanism in The Legend of Zelda is local, as the gate and the method of opening it are in the same location, and could be considered systemic, as the condition is at least consistent in rooms of this type throughout the game. But it is unintuitive, in that there is no intuitive connection between a door being closed and locked, and nearby monsters being alive or dead.
A slightly more complex but common example is a lock & key gate. As in Resident Evil, the Insignia Door may require the Insignia Key to open.
The key is hidden somewhere in the gameworld, and the door will not open until the player possesses the key and returns. This gate is distributed, since the gate and key are in separate locations, special case, since the Insignia Door and Insignia Key are both unique one-off objects, and fairly intuitive, since a door with a unique key is a known concept in the real world.
So let's look at player tool gating. I didn’t work on the first Bioshock, these are only my personal observations, but I believe it contains one of the clearest recent examples of this concept. The second level of the game is Medical Pavilion.
At one point in the critical path, the player must reach the Dandy Dental subarea. However, Dandy Dental is blocked by a broken pipe
which has spilled freezing water into the environment. A thick ice wall blocks the path to the dental wing. This is a gate: the player has few tools available at this point in the game, none of which might clear the blockage.So the player must explore the Eternal Flame Crematorium and gain the Incinerate! Plasmid,
which allows them to shoot fire out of their hand. Returning to the ice wall they are able to melt the blockage and proceed.
This gate is distributed, in that the Plasmid and the blockage are in separate locations, systemic, as all ice in the game can be melted by fire, and intuitive, in that every player knows that fire melts ice in the real world. The primary difference between this type of gating and the other examples is that the key to this gate is a systemic player tool which can be used for purposes other than opening gates. The Incinerate! Plasmid can be deployed arbitrarily at the player's whim, as opposed to the Insignia Key which can only be used in a prescribed location. It can be used to set enemies on fire
conferring damage over time to the target, or to set oil slicks on fire, creating an area of denial that will damage anything that passes through it. It is an ability with a systemic use, which can also open gates, in that order of priority.
Gating tools empower the player because they obey consistent rules of the game, encouraging careful observation and fruitful pattern recognition. Once the player recognizes that fire melts ice blockages-- and placing an ice blockage on the critical path which must be melted to proceed enforces this knowledge-- they are on the lookout for additional ice blockages that may react to their tools in this way. What defines this tool as gatable is one of its systemic properties, which can be internalized by the player and applied throughout the game, allowing the player to recognize opportunities and form their own short-term goals to exploit them. Empowerment stems from this feeling of self-direction. Despite the examples used so far all being literal blocked doors, gating tools are extremely flexible. A gate can be anything that binarily blocks the player's progress. For instance, let's flip the ice example from BioShock around and put the ability to freeze things in the player's hands. This is just what the Ice Beam from Super Metroid does.
The tool's general utility is clear: freezing enemies neutralizes them briefly, allowing you to evade or destroy them while immobilized. But it also lets you stand on enemies while frozen, taking no damage. At the outset of the game, there are certain vertical shafts in the gameworld that are patrolled by small, invincible enemies. The shafts are too high to traverse normally, but once the Ice Beam is acquired, the enemies can be frozen in place and used as platforms to stepladder up to an otherwise unreachable location.
As with any good gating tool, this ability's primary utility is not gating, but systemic play. The blockage is a vertical space that is too high to traverse normally; the player clears it using an emergent outcome of the collision between the properties of a player tool and the properties of an enemy. The player does the mental math, envisions the result, and puts their plan into action. That's empowering.
So a “gate” is not necessarily a gate. It's a body of water you can't cross without a raft. It's a pit of lava you can't pass over without a grappling hook. It's a gas-filled room you can't pass through without a rebreather. or a dark cave system you can't navigate without a flashlight. Given the right tools, the player is empowered to clear these gates in a manner that is distributed, systemic, and intuitive.
But these native player empowering properties of gating tools are only half of the equation. How can their arrangement in the gameworld guide the player down the path authored by the designer? When I was initially designing the Minerva's Den singleplayer expansion for BioShock 2,
I knew that I wanted to make it its own self-contained side story, and condense the player growth schedule of the full game into the DLC's two-and-a-half levels, allowing the player to grow from nothing all the way to the top of the power curve in only three to five hours. One advantage of this is that it allowed me to start he player out as a blank slate, which is the perfect canvas for some good old player tool gating. This method of guiding the player using gating tools became my guide for introducing game mechanics in the order I intended, and telling the story I wanted to tell. I also knew what story that was:
it’s the story of Charles Milton Porter, a brilliant computer scientist who fled the surface world and founded Rapture Central Computing to push computer technology and artificial intelligence further than it had ever gone before, but whose personal obsessions blinded him to betrayal by people he'd trusted. How does this all go together? Minerva's Den is a district that serves as the technology center of the undersea city of Rapture, home to three businesses:
Rapture Central Computing, which houses the mainframe responsible for all the computation in the city, McClendon Robotics, and the Air-Tite Archives, which uses the city’s pneumatic tube system to store people’s valuables. The level is arranged in a hub-and-spoke structure,
(please ignore my bad photoshop hackery here) with a readily identifiable landmark
at the center of the hub-- this statue, which has directionality that the player can orient themselves by. The player's path through this space isn't immediately clear
but I knew I wanted to define it using staged tool gating. At the outset of the level
[here], the player is contacted by Porter, and directed to his office
[here]. To get there, the player must pass through the hub, which connects to RCC, McClendon, and Air-Tite.
Two of these major connections are gated by tools that the player hasn’t acquired yet: McClendon Robotics is locked off by a Door Control Panel that’s out of reach
and the Air-Tite Archives are locked off by a short-circuiting mechanism that requires an electrical jolt to activate.
and the Air-Tite Archives are locked off by a short-circuiting mechanism that requires an electrical jolt to activate.
The entrance to Rapture Central Computing is open and our initial goal is at the end of that spoke. Arriving at our first goal we acquire the Hack Tool, which allows us to manipulate machines from a distance, and learn that we’ll need the Gravity Well Plasmid to complete the level. Some investigation reveals that a dose of Gravity Well can be found in the Air-Tite Archives. So once we complete our first top-level goal-- getting to Porter’s office-- our full progression is set up. We know that we need to get into Air-Tite to get the tool that will complete the level, but that spoke is gated by Electro Bolt, which we don’t have. But now we do have the Hack Tool, meaning we can get into McClendon Robotics.
By exploring McClendon we find Electro Bolt at the end of that spoke, allowing us to retraverse the hub and open Air-Tite using the newly acquired tool
At the terminus of this final spoke we find Gravity Well, which was designed to have a crowd control combat function
as well as a gating function. Returning to the exit of the level, we can now throw the Gravity Well projectile over a barrier, unseating the capacitors in this bank of machinery and disengaging the magnetic locks.
This is the resulting player path that’s drawn through the level.
Each major point where the player hits a terminus and turns back around, we have a guaranteed chokepoint where a new critical element of the player’s toolset is introduced.
This becomes our primary storytelling progression as well. Here we see portraits of the co-founders of Rapture Central Computing, found in the hub of the level. At each of these chokepoints,
I seeded the next major beat of the backstory, setting up the concepts that would lead to complication in the second level and the resolution in the final level. One of the primary methods of delivery was the Audio Diary,
allowing me to place small snippets of story along the path at a pace that complement the gameplay flow.
As the player pieces together their toolset, they’re also piecing together the story of the people and events that made Minerva’s Den what it’s become.
Through this example we can see how we take a city district arranged as hub-and-spoke and weave a reliable path through it, pairing major story beats with the introduction of the player tools that will allow the player to access new areas, maintaining a 1:1 relationship that is as reliable and robust as any strictly linear world arrangement might be. Each element of the progression supports the other, leading, hopefully, to an organic experience of the entire space, which the player feels empowered to pursue and rewarded for completing. This linear path, while inherently unnatural, is presented in a systemically naturalistic way. With any luck, the illusion is maintained and the designer remains invisible. So what are the guiding principles I discovered during this process of building Minerva's Den using stage player tool gating?
Once the player reaches Porter's office, they know that they need to get the Gravity Well Plasmid to complete the level. Their pursuit of this end goal drives them to seek out the tools that will unlock the intermediate gates. “I must get Gravity Well from the Air-Tite Archive, but it's gated by Electro Bolt. I must get Electro Bolt from McClendon Robotics, but it's gated by the Hack Tool.” If the nature of the gate can be determined by the player organically, all the better-- they perceive a prerequisite, and give themselves the goal to fulfill it.
Referring back to the Medical Pavilion example, the ice wall is a clear and intuitive gate, but one that, throughout the game, may be cleared by any source of fire: a burning object carried with Telekinesis, the flamethrower weapon, exploding shotgun shells, or an exploding barrel all work. So prior to the introduction of Incinerate!, there are no other sources of fire in the level. Similarly in Minerva's Den, we had to remove any weapons that did electrical damage, like Electric Trap Spears, from the level prior to entering the Air-Tite Archives. Allowing the player to clear gates without using the intended gating tool leads to sequence breaking, which has tangible knock-on effects in terms of bugfixing and maintaining parallel content to account for achieving goals out of order. Supporting sequence breaking is done in minor ways in Minerva's Den-- due to the open-ended structure of the levels, the player can acquire the final gating tool without being told they needed it in the first place. But this effort allows us to directly acknowledge the player’s cleverness in completing the level out of order. It becomes a memorable moment for those players most interested in defining their own path through the level.
By their nature, the use of player tool gates tends to lead to some amount of backtracking: the player ungates an area, explores it, obtains a new tool, then leaves the area and returns to one that can now be ungated. This results in an A-B-A-C-A-D... rhythm to the level, with A being the hub. This can be a good thing, as it extends the critical path without adding level geometry, allowing the designer to reuse playable space on successive visits by the player. To this end, the gating tool should lie at the far end of the spoke,
and there should be a clear and unobstructed central path leading to it, with optional explorable spaces off the sides. This way the player can fully explore all of an area's side spaces as they work their way toward the gating tool, then easily return via a straight shot back to the hub once they've acquired it.
The first level of Minerva's Den is laid out as a hub with three major spokes: Rapture Central Computing, McClendon Robotics, and the Air-Tite Archives. When I first laid out the player tool gates in the level, I allowed the player to explore about 80% of McClendon Robotics and Air-Tite Archives before encountering the critical path gate that kept them out of the final subarea of each respective spoke. Our Design Director at 2K Marin, Zak McClendon, pointed out my mistake here: one major component of player tool gating is a map function
-- areas that have been visited are filled in while areas that haven’t are grayed out, a sort of fog-of-war effect. The player checks the map to see what they haven't visited yet and by process of elimination clears the available areas to find the means to access the currently gated ones. By letting the player explore the majority of a spoke before encountering a gate that they don't have the tool to clear, their map shows only one or two rooms that remain grayed out in that spoke. This unexplored space seems inconsequential. Anything could be back there-- it could be a broom closet, or the level boss's lair, but its relevance is unclear. Whereas pushing the gates to the entrance of each major spoke
called the blockage out as inherently significant. They were blocking a huge, named section of the map, with a big sign over the door
, and therefore begged to be cleared. Know your end goal. Lay out your level as hub-and-spoke. Gate the spokes exclusively via intentionally-distributed player tools. Taking these guidelines into account, we can extrapolate a highly schematic level sketch defining a reusable framework free of specific content.
We can see here the hub & spoke structure, the A-B-A-C structure, and the way that the player would organically clear the map through exploration and process of elimination. This could be expanded to a theoretically endless chain...
But this highlights one of the inherent limiters of this approach, which is growing complexity. The required tools form a nested goal structure-- “I must get A to get B to get C to get D to get E to get...” which before long turns into an unmanageable run-on sentence. We never want the player to lose sight of their end goal which, when buried under too many nested subgoals, can easily be forgotten altogether. “Why the hell am I doing all this?” is never a question we want the player asking themselves. A more sustainable expansion would be a series of hubs each separated by a gate:
Each hub contains a new goal, requiring the player to traverse the spokes, working their way down the path you've authored via the placement of gating tools. You'll note that these last two diagrams both present the same number of gating tools, but it's clear which one is more parsable, but staging the hubs the same way you'd stage the gates and tools. So personally, I consider this a solved problem. It's about as close to a scientific formula as you can get when it comes to level structure. It's reliable, sustainable, and extensible. It's just waiting to be filled with design content. We've exposed the formula, and know what problems it solves. Of course, it's not a one-size-fits-all solution. So let's talk limitations.
One limitation falls out of just that formulaic nature:if we're trying to avoid a feeling of artificiality, exclusive use of this approach on a large scale can become as transparent as any other gating mechanic. Super Metroid is probably the purest example of the formula being applied as the exclusive gating method throughout the game, but generally in higher fidelity games, moderation is the key to success.
Batman: Arkham Asylum is an outstanding example. At key points throughout the game, the player gains new gating tools such as the Grappling Hook, Explosive Gel, and the Line Launcher.
But between these player tool gating points lie story-based gating objectives such as saving hostages or interrogating villains. In a game with story and characters, variety is important to making the experience feel textured and alive.
That said, the degree to which you can apply this formula is limited on a lower level by the available player tools. This gating method is only used in one level of Minerva's Den because there are only a handful of gating player tools available in those games. It's intuitive that Incinerate! can clear ice, but how does one gate on a swarm of bees?
Or Enrage? The game doesn't include traversal-based tools such as a double jump or a grappling hook. The available palette isn’t huge.
But even with only a small number of gating tools in the game, why not spread those tools out more judiciously over the full course of the game, extending their usefulness across more levels? The answer points to another limitation of player tool gating: its impact on longterm player choice. Why do you gain most of your tools in the first hours of BioShock 2 or Minerva’s Den? And why does Deus Ex Human Revolution offer a multitude of player tools that would be ideal for crit path gating, but never structure its levels around an authored tool gating progression? Because these games emphasize longterm choice based on these player tools themselves in ways that Metroid and Arkham Asylum don't.
Having access to a wide variety of player abilities allows the player to define a playstyle in a subtractive manner. Do you want to be the guy who focuses on stealth and melee attacks, sacrificing raw firepower and hitpoints? Do you want to be the player that approaches combats indirectly, by hacking turrets and cameras and turning enemies against one another? Or do you want to crank up the damage output of your direct offensive abilities and go full aggro? This emphasis on tailoring a playstyle using a small subset of the many abilities available, and on upgrading individual abilities over time, requires as many of the tools as possible be doled out early, to give the player that broad palette to choose from and refine. Holding off the gating tools over the course of the game would work against these higher-level goals, as the player wouldn't have as many options to experiment with up front, and wouldn't have time to upgrade tools that were introduced late in the campaign. Player tool gating in Minerva’s Den was more a technique of opportunity than a focus of the overarching design, one that is only present in the first part of the experience. Or take Deus Ex Human Revolution.
One of the publicly-stated design principles of that series is “multiple solutions to every problem.” Players are intended to build their character the way they want, and feel like they've invested wisely whenever they approach a goal: there's a path for the hacker, a path for the environmental resistance guy, a path for the high jump guy or the super strength guy. And these tools can be acquired in almost arbitrary order; after the first few hours, it's impossible for the designers to know whether a player may have maxed out their hacking or invisibility or jumping & sprinting abilities. To define a single path using player gating tools, there would have to be critical path blockages that supported “one solution to one problem,” and the gating tool would need to be placed in a specific location in the gameworld by the designer. Again, the limitations of tool gated progression would work directly against these core design goals. Tool gated progression is deterministic.
Your Samus at the end of Super Metroid will be essentially identical to mine; your Batman at the end of Arkham Asylum is only marginally different than anyone else's. Whereas my build at the end of BioShock 2 or Deus Ex might be wildly different from yours. Is it possible to strike a finer balance between player choice and a mandatory tool introduction schedule? Sure, but the core limitation remains: a gating tool is a guaranteed addition to the player's experience.
Will you have the Insect Swarm or Hypnotize Plasmid by the time you finish Minerva’s Den? Who knows, but you’ll definitely have the Hack Tool, Electro Bolt and Gravity Well. It's a trade-off. Is it cool that I can be a master hacker within the first few hours of Deus Ex Human Revolution, or complete the entire campaign without upgrading hacking once? Personally I'd say yes, but gives the designer one less tool the designer has for crafting an intentional arc of player abilities and ordered access to specific spaces throughout the game. Know what you're gaining, and what you're giving up.
On a different axis, your available tools are limited by your fiction. Metroid is an outer space sci-fi story; BioShock 2 includes fantastical genetic mutations; Deus Ex is based on superhuman cybernetic augmentations; Batman has all his gadgets. If your fiction is based around, say, the Vietnam War or a 40's police investigation, it's going to be harder to find a place for double-jumping or freeze rays. But gating tools can also be low tech:
the ax example is put into practice in Condemned: Criminal Origins by Monolith.
it’s a game about melee combat using improvised weapons, some of which are used for gating.
The fire axe is the only tool that can break down cracked doors, but it’s also used as a heavy-hitting weapon in combat.
There’s also a crowbar that can be used as a bludgeon, and that can also pry open safes and metal doors which would otherwise be impassable. Don't discount your ability as a designer to spec expressive gating player tools, even if they may not be flashy or appear in great numbers.
As with any design decision, choosing whether to pursue player tool gating comes down to determining what core values you intend to support. Do you want to present a contiguous, interconnected space that is gradually unlocked section-by-section? Do you value an authored player power curve, or want to tell a linear story within a nonlinear space? Do you want to remain hands-off as a guide and encourage the player to find their own way, using their powers of deduction? To these ends, gating progression using player tools can be incredibly powerful and extensible. It expresses, I believe, a trust in your player, and yourself as a designer-- that you'll be able to arrange the elements of the gameworld in a way that speaks for itself, and that the player will be able to follow these cues to find their own way along the critical path in response. Ideally, I think of game design as a collaborative enterprise between the designer and the player; not the designer giving the player direct orders and the player dutifully carrying them out, but both parties participating in a sort of call and response. The designer fills the gameworld with objects that hold semiotic meaning; the player decodes these cues and responds as they deem appropriate. It's another version of the old axiom, “show, don't tell.” You could tell the player, “find a hammer,” or you could show them a nail. As with most productions, your approach will likely be a mix of the above-- both one-off, story-based goals, and systemic, rules-based goals. But hopefully this talk has provided you another useful method of guiding your players, more invisibly.