Saturday, October 27, 2012

Dishonored: The Missteps

Dishonored has been one of the biggest surprises of this year for me.  Developed by Arkane Studios, of Arx Fatalis and Dark Messiah of Might & Magic fame (or at least, fame in my book), and combining their penchant for first-person melee combat gameplay with Harvey Smith's Deus Ex-ian talents, I was absolutely certain I was going to find myself disappointed in the most crushing way imaginable.  After BioShock set itself up as a spiritual successor to System Shock 2 but ended up being Doom with some pseudo-philosophical jargon added in, I was fully expecting the next big shooter-RPG hybrid to be a let-down.

Yet, it wasn't. Dishonored is a fantastic game, and I had a huge amount of fun playing through it over the last couple of weeks.  Broadly, it was just about everything I wanted out of a game... or at least, about as close as it could be to one not modeled from the substance of my deepest desires and dipped in chocolate.  Artistic, smartly written, with well-defined and well-explored characters, excellent level design that offers options without pigeon-holing players, and atmosphere and ambiance up to its eyeballs, it's a great accomplishment and a joy for me to play.

But, it is my nature to nitpick everything, chronically so, and so while I write a longer article discussing both what I think are the game's successes and failures in more extensive detail, I'd like to discuss what I feel are three major problems that Dishonored has, and what it could have done to improve upon them.


Dishonored takes heavily from Deus Ex, but the second big influence on the game is undoubtedly ThiefDishonored is generally far more focused on stealthy gameplay than it is on Deus Ex's more open-ended, hub exploration model, and the mission structure with objective summaries confirms this as well.  But as much as Dishonored calls itself "inspired by" Thief, the fact is that its stealth gameplay is significantly lacking compared to that title.

In terms of basic mechanics, Dishonored simply does not have as much to offer.  Although light and shadow apparently do have an influence on whether enemies can detect you as you sneak around, I found it to be an incredibly subtle distinction, as enemies were able to spot me hulking in the shadows from distances of 40-50 feet near the endgame.  There is also almost no reliable way to manipulate light and shadow in the game, which was Thief's big technical innovation in 1997... if Thief could do this with dynamic lightmaps running on antiquated technology, why can't Unreal Engine 3.x.x manage the same?  Sure, it's more a design problem than a technical one now, but even so, one would expect that after so many years improving graphics, we'd start using them more consistently to enhance gameplay as well.

Additionally, the lack of tools in stealth is a big misstep.  Thief gave players a wide variety of tools, from flash bombs, to moss arrows, to grappling hooks, to climbing gloves, to water arrows, and more.  Deus Ex didn't offer quite the same selection, but even it had the courtesy to offer non-lethal alternatives like the stun prod and gas grenades.  Dishonored, even with its Blink power that lets players teleport around the environment quickly, feels like it's missing several critical stealth tools with its paltry poison bolts for the crossbow and the ever-popular stealth takedown.  There is almost no evolution of stealth gameplay over the course of the game, which is a real shame when it's apparently the focus of the title.

Last, the way the AI is set up leaves something to be desired.  In addition to being very poor at looking just a few degrees up (the guards in Dishonored's world must play a lot of console shooters), whenever a single guard spots you, every other guard in the area will be alerted to your presence and will magically know exactly where you are.  This means that being detected always forces you into a straight-up firefight, and running away is strongly discouraged in favor of save-scumming.  Combined with the lack of progression and options in non-lethal equipment to use, it means that players rarely have any adequate escape route or options to get out of a fight other than slaughtering the alerted enemies.

Silent Protagonist

Dishonored has a silent protagonist, which many people will point to as an antiquated thing in 2012.  I don't quite agree with this perspective, simply because it assumes that a silent protagonist is a technological or budgetary limitation and not a design decision with upsides and downsides.  However, Dishonored's choice to use a silent protagonist ultimately harms its narrative delivery and the resonance of its themes and characters.

Silent protagonists are usually selected because they allow players to project onto them, serving more literally as the player avatar in the game world.  However, Dishonored has a very defined character already - Corvo Attano.  Corvo has a definite origin, upbringing, occupation, history, and past relationships with key characters in the game.  He's not some newbie punk who's working his way up from the bottom - he's one of the most important people in the empire.  With such an established character, it becomes hard for players to project themselves into the game - so the other approach of creating a character that players identify with is almost required.

In Deus Ex, JC Denton was not a great character, but rather than an avatar, he was someone the player could identify with in the third person.  I never felt like I wanted to be *me* in Deus Ex - JC was appropriate to the setting and themes of the game.  It let players learn his motivations, his history, and then make choices based on how they interpreted the character.

Dishonored attempts involved conversations featuring the player, but Corvo never utters a word, so in addition to just being a bit weird when more extensive dialogue sequences occur, it's very difficult to get a sense for his true motivations, his relationships with other characters, his personality.  Additionally, many sequences feel like they move too quickly because of the lack of dialogue on his part, and some plot holes could have easily been filled in if Corvo could simply ask questions and direct the discussion himself.

Character Progression

I think that Deus Ex's tri-tier character progression system of weapons and equipment, skill points, and augmentations was a work of genius.  Each one of these systems grew and developed over the course of the game, each had unique qualities which were rarely redundant with the others, they were all rewarded through different kinds of gameplay and exploration, and most importantly, they offered opportunities each for all kinds of play-styles, not just one or two.  Sniper?  There are augmentations, weapons, mods and skills for that, just as there are for a demolitions expert, a melee assassin, and more.

Dishonored tries to do similar, but is far less successful due to poor pacing and a lack of options.  Bone charms are effectively equippable passive perks found by exploring the environment, and give small bonuses such as extra health or quieter movement.  However, they are distributed randomly around the game world, so their capabilities had to be fairly generic and limited, since a player could get any of them at any time.  It also makes planning character "builds" for replays difficult.

Runes, meanwhile, are currency used to unlock new occult powers and more substantial passive upgrades, similar to augmentations in the original Deus Ex.  However, the lack of significant upgrade tiers for individual powers, and the relatively low number of them, means that players will have purchased almost all the powers they need only one or two missions into the game.  Additionally, the ability to buy any power at any time means that it was impossible for the level designers to create really interesting, exclusive challenges and solutions - sure, you can sneak in through a drain by possessing a fish, or summon a rat swarm to devour your enemies, but why bother when you can just use an equally-effective grenade, or the deviously powerful "walk through front door" ability?  There's always a fool-proof solution, and it's usually easier than the one requiring you use a power anyway.

Last, weapons can be upgraded and modified.  There is a distinct lack of weaponry throughout the game, especially for stealthy players, and while I appreciate the attempts to make limited equipment feel more valuable, the fact is that there is very little "heavy artillery" for more action-oriented players.  Additionally, blueprints can be found and bought throughout the game, but all these do is upgrade the ammo capacity and effectiveness of your weapons in a completely linear way - there are no trade-offs to make.  And, like the powers, weapons really don't get more interesting or varied as the game goes on, leading to a feeling of stagnation around halfway through.  Again, it's nice to see an action game that doesn't thrust a minigun and bazooka in your hands, but played as a shooter, Dishonored doesn't get very far because, like the stealth aspect, it has significantly less to offer players than its inspirations.

Closing Thoughts

As I've said, I still enjoyed Dishonored quite a lot, and I'll be discussing its successes and failures in more detail in a follow-up.  It's also worth noting that not all of these is an example of "bad design" - but rather, they're things that I personally believe are inferior to what could have been.  I really do have to wonder at the justification regarding some of these decisions - did it come down to time and budget, or ease of development, or a more conscious choice to limit the player's capabilities to encourage other elements of gameplay?

Friday, October 12, 2012

Musings on Successful Multiplayer Maps

If you're a fan of multiplayer gaming, especially shooters, chances are you can recite the layout of a few classic maps flawlessly.  Whether you've played Halo, Counter-Strike, Call of Duty 4 or anything else, there's probably one or two maps in those games that stand out for you as being particularly fun, replayable, or, for lack of a better term, memorable.  But what sets some maps apart from others?  How do we get "fun" out of one layout, even when it might be similar to another one?

In this article I'll be turning my attention to what I think are the core components of an interesting, fun, replayable multiplayer map, and how these differentiate a good level from a great level.  For the sake of simplicity, this article is primarily concerned with competitive multilayer shooters, but I hope some of these ideas can be applied to other types of games as well.

Cyclical Flow

Mihály Csíkszentmihályi's concept of psychological flow is something that has been oft-studied by game designers in order to help create games that are more consistently engaging for players.  The basic idea of flow (forgive the over-simplification) is that it is a mental state individuals become immersed in, wherein all their attention, emotions, and physical activity are concentrated on a single task; unlike "typical" work, where we might find ourselves following lists in order to complete tasks, flow allows for one task to move naturally to the next, to the point where we don't even stop to consider that it is even happening at all.

In the field of level design, flow can be expressed as a state players will find themselves in wherein they move naturally from one task, point of conflict, or objective to the next, in a smooth and resistance-free way.  If you've ever played on a particular multiplayer level and found that you end up moving from one place to the next along with all other players, and find that flashpoints of conflict and contention arise without you necessarily "working" to bring them about.  Consider too, how many "bad" multiplayer maps are also ones where sometimes it's not intuitively clear where to go, or what to do, and the action often seems to pass you by.

A good level design will be able to create this state in players nearly all of the time.  Generally, this is accomplished by using cyclical layouts - that is, levels which direct players in loops which circle back in on themselves and where the natural movement of players leads them to constantly run into objectives, opposition, etc.  The easiest way to demonstrate this is to show it directly:

Apologies for my poor penmanship!

Here's a map that almost any shooter fan should be familiar with: Counter-Strike's de_dust (or just "Dust").  Above, the red lines and "T" mark represent the routes and starting position of Terrorist team players, while the blue lines and "C" mark represent the same for the Counter-Terrorist team.  If you examine the level, it becomes clear that the routes it presents to players converge, roughly concurrently, at several key conflict zones throughout the level, denoted with yellow "X" marks.

There are a lot of nuances to this simple but effective layout, visible simply by following the routes above.  For instance, if the Counter-Terrorist team decides to slow down and play defensively, this will naturally draw the Terrorist team to the bomb site on the east side of the level and the center area.  If Counter-Terrorist players rush the Terrorists all-out, the conflict zones will transition to the tunnels in the center of the map, where three separate routes meet each other.  If the Counter-Terrorists decide to take the west underpass route, and no Terrorist players decided to go that way, they will eventually loop back around on the Terrorist players, giving both a flanking opportunity to the Counter-Terrorists and a chance for the Terrorists to take the objectives at either the east or south sides of the level, as well as to prepare for the inevitable assault.

It's important to note that the amount of time it takes for each team to traverse a given portion of the level is almost exactly the same as what it takes the other team to traverse their own.  This means that players, even if they run aimlessly around the level, will almost always encounter an enemy eventually, usually within 30-odd seconds.  When the players have been killed off on both teams and a 1v1 situation as arisen, the need for the remaining Terrorist to plant the bomb at one of the bomb sites means that the Counter-Terrorist now has an ultimatum to finish the round, but can also immediately know that the Terrorist is hiding out at one of two locations.

Despite the fact that Dust is a fairly simple map by some standards, it has endured because the lack of distractions, redundant routes and other level design "clutter" allows players to more easily achieve a flow state, and when they have gone beyond that first 30 seconds of gameplay each round, the design of the map ensures that players will cycle around it naturally until the end.


If there's something in common with just about every great multiplayer map I've played, it's that they all have some sort of vertical component to gameplay - that is, the level does not provide one flat plane to run around upon, but has two or three different ones which intersect with each other in interesting ways, or provide new gameplay opportunities.

Verticality does two important things.  From the perspective of the player, it makes combat more interesting and dynamic because enemies and friends can come not just from left and right, but from above and below as well.  It taps into that 360-degree perspective players have of the world in most shooters, and creates more interesting tactical advantages.  For example, something as simple as being able to use high ground to fire down on an enemy from cover, or the opposite, throwing a grenade from below into an upper window to flush an enemy out, can present opportunities for players which stray from the usual "aim -> shoot" flow of gameplay.  What's more, these are often decisions that need to be made in a split second, as the risk/reward element inherent in them means that there are multiple choices presented to players beyond the most basic ones.

Second, verticality also allows for more interesting routes between levels.  In the above example, I demonstrated 4 main routes that the two player teams each can take before meeting each other, but what is also apparent upon examining these routes is that they often overlap or intersect with one another.  For instance, the route leading under the west overpass can be flanked from above if a player takes the tunnel route through the center of the level and heads west out on top of the overpass.  The westernmost Counter-Terrorist route, meanwhile, leads them directly to a ledge overlooking the Terrorists' route under the overpass, allowing for a crafty player to set up a point defense against them.

Of course, de_dust isn't the only level out there featuring vertical gameplay.  Another excellent example of this can be seen in the much more recent "modern classic" Call of Duty 4 map, Crash.

Yellow X marks denote buildings with multiple floors and windows.
Crash is not specifically designed as a vertically-oriented level.  However, it has a few vertical elements which add up to make it much less interesting than it would be if it simply took place on a flat plane.  Above, the buildings with yellow X marks overlook key parts of the map, in this case the center courtyard with a crashed helicopter, and some back alleys that serve as an entrance or escape route from the main courtyard.  Almost every single major route, except for a few transitional routes between key areas, is covered by a two-story building with a window.

What this means is that players have to constantly be on the lookout for enemies on the ground level, but in buildings as well.  The key conflict zone is, of course, the central courtyard, and it has no less than three buildings providing partial coverage of it from different angles.  This three-way split means that no single vantage point has easy access or line of sight to all the others, creating a natural "rock, paper, scissors" back-and-forth relationship as players navigate the level.  As all routes funnel into the courtyard, it is almost impossible to not end up in it eventually as players try to get from A to B, with the only alternative being to take the back alley route, which often breaks out in skirmishes as players try to flank each other inside the courtyard.

The tactical importance of these buildings hinges almost entirely on the fact that they provide cover and a more complete line of sight across the battlefield, not to mention potential sniper spots.  However, defense of the buildings also becomes a crucial element in deciding the winning team, as suddenly, weapons like claymore mines become useful in stopping other players from storming a building and attacking its occupants from behind.  They are important because of the vertical element they introduce to the level, and that importance carries with it many other secondary but equally critical elements.  It all adds up to make for a very entertaining, replayable map which constantly evolves as players try to take key points from each other and move on to the next.

Strategic Distinctiveness

The third major component that goes into a memorable multiplayer map is a bit harder to pin down, but in my experience, it boils down to something I'll call strategic distinctiveness - certain opportunities that players can use and exploit on a level which make it very different to play on than anything else.  Often these will be near-ritualistic behaviors players perform over and over every time the map begins, while other times they are simply little tricks that more experienced players can use to their advantage.

Below is a list of just a few examples of these, but almost every multiplayer map that keeps players coming back has at least a couple of these sorts of qualities:
  1. In de_dust, players who get very good at timing grenade throws can almost single-handedly wipe out the other team as they stream in through the central tunnels.  By estimating the enemy team's speed and using sound rather than sight, it's possible to negate the other team's advantage and destroy their body armor or take a big chunk out of their health with one well-timed grenade.
  2. In another Counter-Strike map, de_dust2, a large pair of double doors, open just a crack, overlook one side of the Counter-Terrorist spawn exit.  A caffeine-fueled Terrorist player with a sniper rifle can, with a lot of skill and cat-like reflexes, take out several Counter-Terrorists each round if they are not careful to leave their spawn.  This also encourages them to use alternate routes rather than just rushing to the closest bomb site, preventing the Counter-Terrorists from becoming prematurely entrenched on one side of the map.
  3. Team Fortress 2's ctf_2fort map allows players of various classes to exploit the environment to their advantage.  For example, Soldiers can rocket-jump to the enemy team's battlements from the central courtyard, allowing them to wreak havoc on the snipers that usually rest there.  Pyros and spies can also learn to lurk around key zones, including the paths leading to each team's flag/briefcase room, or the lower tunnels leading in from the courtyard's moat.  Snipers, of course, can use the battlements as a vantage point to prevent the other team from crossing the courtyard, making counter-sniper team tactics essential.
  4. Call of Duty 4's Broadcast level features a large newsroom-type area featuring dozens of desks.  The control booth sitting above makes for an excellent nest for machine-gunners and snipers alike, and often holding this point can ensure a steady stream of kills.  The well-defended position gives ample viewing angle across the entire room, but is consequently prone to grenades tossed from below or players sneaking up from behind.
This doesn't mean, of course, that every multiplayer map has to look, feel and play completely differently from all others.  In fact, many multiplayer maps resemble each other, in terms of layout, very closely, and tend to only differ in the smaller details and, of course, visual style.  However, just one or two room configurations, hiding places or other little elements that no other level out there has can really make one leave its mark on players.

Closing Thoughts

Of course, the science of what makes a memorable multiplayer level is still subject to a lot of discussion and interpretation.  Certainly, developers have created games in the past whose "breakout" maps have been ones they spent the least amount of time on, and there always be some maps which defy all academic notions of quality.  In fact, some levels, such as Counter-Strike's fan-made and eternally popular fy_iceworld, are about as close to "bad" as you could possibly get by almost any level design standard, and are perhaps proof that to some degree, success in creating something players will remember for years to come is not something which can always be deliberately engineered.