Gamasutra: The Art & Business of Making Gamesspacer
Why Left 4 Dead Works
View All     RSS
August 30, 2014
arrowPress Releases
August 30, 2014
PR Newswire
View All

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

Why Left 4 Dead Works

December 21, 2010 Article Start Page 1 of 4 Next

[What makes the Left 4 Dead series' gameplay design work so well for players? This comprehensive design analysis analysis from Full Sail grad student Paul Goodman and course director Adams Greenwood-Ericksen takes developer commentary and matches it against the SRK Framework, a tool for categorizing cognitive behavior in interactive media.]

Mechanics are the core building blocks of any game. They are the primary method by which players are able to operate in a game environment to overcome the challenges and obstacles that a game's developers have placed down. Yet in their design a crucial factor can be overlooked with regards to how game mechanics are made or adjusted through the development process: understanding the ways in which players think and learn about the function of game mechanics is key to determining their effectiveness.

If a mechanic is too complex for players to understand, the result may be player frustration, whereas if a mechanic is too simple, it can quickly lead to player boredom and repetitive gameplay. By using established methods of assessing and modeling cognitive behavior, game designers might gain better insight into the likely thought processes of players, enabling them to improve the quality of mechanics in game.

One such method, the Skills, Rules, and Knowledge Framework (SRK), which was first presented by Jens Rasmussen in 1983, focuses on categorizing different types of cognitive behavior across three different levels.

  • Skill based behaviors are activities that are common and routine for an experienced individual and do not require much attention to execute; instead, the process is "triggered" and carried out in an automated fashion. An example of a skill based behavior common to console gamers is using the left and right thumbsticks to look around and move an avatar in game.
  • Rules based behaviors hinge on following known mental steps and checklists built to guide the activity and require individuals to put more effort into carrying them out. Preparing to load a saved game on a console is be an example of a rules-based behavior, as it follows a checklist of steps to navigate through menus to find the appropriate saved game file and load it.
  • Knowledge based behaviors present players with an activity with which they have little or no prior experience or understanding, and therefore must undertake some form of research (e.g. reading manuals or checking forums) or problem solving (e.g. trial and error experimentation) oriented approach to gain more information. A longtime gamer who's never played any form of major MMORPG, for example, would very likely engage in knowledge-based behavior in order to understand the functions of a typical interface from that particular game genre.

A strength of the SRK framework is that it is broadly applicable to virtually all interactive activities. Therefore, in the context of electronic gaming, players of a particular video game will engage its mechanics using cognitive behaviors that can be assessed and classified by designers using the SRK Framework.

Therefore, game designers can use this tool to identify how players engage a game mechanic during gameplay. Through reviewing how players move from a knowledge based behavior to rule based or skill based behaviors (or vice versa), designers can assess the effectiveness of a game mechanic and make adjustments as necessary, such as simplifying or adding complexity to the mechanic itself.

An example of a game that was developed with clear emphasis on what and how players think can be found in Valve's Left 4 Dead, released in 2008, and its sequel Left 4 Dead 2, released a year later in 2009. An action/survival first person shooter, L4D puts the players in control of four survivors battling their way to safety during a zombie apocalypse.

Hordes of "Infected" bar their path at every turn throughout the four campaigns, and working together as a team is often the only way players are able to achieve success, especially as many of the challenges posed by game mechanics and obstacles players face require more than one individual to complete.

Left 4 Dead

Early design of Left 4 Dead came about in 2004 after the release of Valve's classic tactical shooter Counter-Strike: Source. According to L4D team member Mike Booth, while experimenting with the AI for non-player bots the dev team found that the concept of "Small team of friends against hordes of clawing enemies" had a great deal of potential. Through more playtests and design discussion, the framework was laid down for Left 4 Dead core concept of a "Co-op vs. the horde game."

Players of L4D find themselves battling against the Infected, victims of a quickly and highly contagious plague that's ravaging the U.S. At their disposal, players start off with a pistol (with unlimited ammo) and a default melee attack for pushing enemies away; throughout each campaign players find more weapons ranging from pump-action shotguns to fully automatic assault rifles, as well as improvised items in the forms of Molotov cocktails and pipe bombs.

Common Infected are the most frequently encountered enemy type within the game, located almost everywhere in each campaign. They appear as normal people, albeit with glowing eyes and bleached skin. The average Infected is a fast runner, can move through most terrain and over most obstacles and assault players with punches and kicks. However, they die fairly easily with a few well placed shots to their chest, head, legs or arms (the horror movie standard of zombies dying only from head wounds having been thrown out).

They also are lacking any sense of self-preservation, and will run through and into hazards and obstacles, including fires and live explosive devices. Occasionally, common Infected attack in large groups (referred to by players as a "Horde") and try and overwhelm the players using their superior numbers. This can be a random occurrence or triggered by a scripted "crescendo" event, usually caused by players having to overcome an environmental obstacle such as a lowering a bridge or destroying a barricade.

The common Infected serves a variety of different purposes, the most obvious of which is to serve as the main enemy type opposing players, but also to provide experience with core gameplay dynamics such as the use of the weapons and items at the player's disposal.

The simplicity of the common Infected allow new players to L4D to quickly evolve during their normal progress through the game environment from knowledge-based behaviors involving research and experimentation intended to determine how common Infected act, to using rule or skill-based behaviors to quickly and efficiently overcome even large numbers of these basic enemies.

Article Start Page 1 of 4 Next

Related Jobs

AtomJack — Seattle, Washington, United States

Level Designer
GREE International
GREE International — Vancouver, British Columbia, Canada

Senior Game Designer
SAE Institute
SAE Institute — San Jose, California, United States

User Interface Design Instructor
SAE Institute
SAE Institute — San Jose, California, United States

Compositing Instructor


Dave Sodee
profile image
Love L4D and L4D 2

Ben Lewis-Evans
profile image
Nice use of the solid examples provided by the L4D franchise to highlight the different levels found in the SRK framework.

It may not be a problem for the L4D franchise, but do you think this kind of "building on the skills and rules of past games" approach to making sequels risks potentially alienating new players? The skill type behaviours in particular can take a long time to form, so it is possible that a new player picking up a sequel could find it hard to cope with the added difficulty of the new additions to the game aimed at challenging older more experienced players.

Perhaps an example of this is could be the online component of Starcraft 2? Given the popularity of Starcraft as a multiplayer game there were a lot of people out there with many of the automatic Skills and Rules needed to successfully pick up Starcraft 2 and be quite competitive right off the bat. Anecdotally that the fact that so much skill transfer was possible for experienced players was intimidating for novices, or people that had played Starcraft only when it originally came out and not kept up with it - This meant that many of my friends who brought Starcraft 2 completely skipped the experience simply because it seemed (rightly or wrongly) like experienced players had too much of an advantage.

Finally, a bit of shameless self promotion. I did a blog post on Gamasutra about the SRK framework and gameplay a week or so ago. It was more of a general overview of the framework, and therefore is less focused than this article (so maybe less useful), but if people are interested they can check it out here:

Matt Hackett
profile image
Good stuff, but not much we haven't already seen and heard for ourselves in the in-game developer commentary :)

Mark Venturelli
profile image
The entire first section and conclusion had me frowning all the way. Also, the analysis makes some good points but is still a long way from really describing the main factors that make L4D (2 more than 1) such an effective design.

This industry hardly needs another buzz-word or work process, and the use of the SKR Framework strikes me as completely derivative of what is already at the core of the game design activity. Every designer's goal should be to provide players with meaningful decisions. In that case, the "meaningful" part is minimizing the "rule" behaviour, and the "decisions" part is minimizing the "skill" behaviour. Players should be making decisions, not using patterns that they have already mastered. When this happens, the game has been resolved and becomes boring (also note that I use "skill" in the context of the SKR Framework, not in a "physical skill" sense - people still make decisions while shooting people on multiplayer FPSs).

Also, every paragraph that contains the word "mechanic" was written in a way that made my stomach turn. "Players engaging a mechanic" was one of the main offenders. As usability/user experience researchers, you should be more aware that in a proper game system players always encounter dynamics, not mechanics. Also "complexity", in the context of this article, should *never* be a design goal. You are talking about depth, which is a completely different thing. Standard MMO combat is very complex, with a bunch of skills and cooldown and effects and what-not. But it is shallow as all hell, usually having the player employing a memorized pattern of skills (that offer the most efficient combat results possible, making other combinations futile) over and over again, against predictable enemy behaviours.

Wayne Wang
profile image
Love L4D

Rob Allegretti
profile image
I'm sure I'm in the minority, but I absolutely hated L4D. It got very old, very fast - the gameplay was pretty much the same no matter the version or level you were playing. Also, I want people to start coming up with better more thought-out enemies than hordes of greedy-AI zombies.

Zombies are like the easy way out - they require no backstory or introduction, and they don't have to talk or display any human qualities. Maybe the success of L4D is indicating that the general game-consuming population is becoming so dumbed-down that this is exactly what they want instead of an intriguing story with well-developed characters.

The whole of the game is just too generic and mindless.

Nicholas Melillo
profile image
I completely agree. Not to mention both games were repetitive as all hell, with both player dynamics and storyline. It's not even the zombies that make the game rather dull, even if the human race had a back story on how they became zombies, what the outbreak was, even the stories of the characters. Maybe throw some flashbacks of the characters memories of their lives in the game, allowing you to control them in the normal world without zombies. How they met, where they come from or even how they survived the contamination. There is just so much more that could have been touched on to make the game more interesting and less mindless.

warren blyth
profile image
I'm at the quite the opposite end of the spectrum (loved L4D and played it to death).

Two quick points:

1) the gameplay was actually pretty much DIFFERENT every time, no matter the version or level you were playing - as long as you played with different strangers online.

I think this is the core of the game's success. it was a game about figuring out how 3 other people were going to play it with you. Core to the whole concept.

(serious. I enjoy playing with pros, because I have to focus on not making mistakes and try to keep up with them. I enjoying playing with bad/noob players, because they are often more vocal/excitable and harder to keep alive.

I still am suprised by the weird ways in which people "manage" their involvement in "the team" - which is a requirement of the game)

2) Storywise, I don't see how they could have integrated the lenthy dialogue scenes of Mass Effect or most RPGs. The story came through quietly in the graffiti, environment design, and dialogue.

There was actually a lot of writing in the dialogue, but you couldn't scratch the surface of it on just a couple play throughs. The game rewarded replay by dropping different major lines for specific times/actions/situations. I didn't appreciate this until I'd played the game for around 80 hours, and then heard Louis ramble on about his youth in the boy scouts for the first time, ending with a cracking about worthless it all was now - which made me laugh out loud.

The second game also, explained a bit of the backstory for the nature of the disease, but people didn't seem to pick up on it. I think this was mainly because of quiet way in which they laid it out - without rubbing your face in it.

just trying to say : L4D franchise takes several non-standard approaches to story and replayability.

E Whiting
profile image
I'm glad someone else agrees. Left 4 Dead is not so much an example of good design, but an example of a game which succeeds in spite of its design.

The campaign gameplay is very simple, offers no depth, and is the same from map 1 to map 20. Zombies run at you, you either shoot them or shove them away. You've got 3 types of weapons to choose from: Rapid Fire (Uzis, M16s, etc.), Burst Damage (Shotguns), or Sniping (Single-shot rifles).

In L4D1 the gameplay was particularly bad given literally every enemy but the tank could be defeated by standing in a corner and shoving (L4D1 has no melee cooldown). The same problem returns in L4D2 once you get your hands on a melee weapon. In L4D2 you can theoretically be hurt by the Spitter or Charger when doing this, but realistically the enemy AI is not intelligent enough to challenge competent players.

The only real improvement L4D2 had was to make level-events where players must progress while the horde attacks. However, one major flaw that exists in both L4D1 and L4D2 is that players who move quickly can outrun the zombies and take advantage of the engine limitations in only allowing 30 zombies at a time - In other words, if you're quick enough, if you can run through the initial areas without killing many zombies, you'll be virtually unopposed for the rest of the map. L4D2's "moving crescendos" reinforce that this is the best strategy in the game, since you're forced into doing it.

The fact that rushing is the best strategy also belies the claims made in the article, L4D's difficulty is actually punishing for new players, but easy for experienced players. New players are most likely to wander around the level, exploring and looking at things. They are also the most likely to be hurt, moving slowly or with extreme caution. In L4D the faster you move through a level, the easier it is. The slower you move through a level, the more common and special infected enemies you will have to face. In other words, the worse you are at the game the more challenges it will throw at you, whereas if you are good you can easily outpace the game's threats. This is exactly backwards of how the game is intended to behave, but it's specifically the attempts to adjust the game difficulty that result in this perverse scenario.

The so-called AI director is ultimately a failure at providing anything more than the most superficial differences in gameplay between playthroughs. Play a map 2-3 times and you will quickly recognize the locations items can spawn in. Yes, occasionally a random horde spawns in a different place, or the pattern of zombies milling about a location will be different -- Games have had semi-randomized pickup locations and semi-randomized spawn locations since Doom (or earlier), so this is hardly an exciting new development.

The actual meat of the L4D games is in the multiplayer Versus mode ... I currently have over 1200hrs of time playing L4D1 and L4D2 Versus, probably more time than the designers themselves. And while Versus can be fun, it's essentially unplayable unless you use modified game settings (Confogl).

L4D1's Versus mode has a distance+health-based scoring system that perversely encourages survivors not to use healing items they possess. L4D2's Versus mode scoring is superior, being based solely on distance+surviving until the end.

However the main flaw with both Versus modes is that the team which "scores," the Survivors, is almost guaranteed to complete the level. Every game between competent players is like a game of baseball where both teams score 10 runs per inning and the game is won by a tiebreaker based on how quickly the bases were run.

Additionally, while this article reveals that plenty of attention was paid to the Survivor experience of play, the Special Infected experience of play is quite unsatisfactory. Bugs and frustration abounds, mostly due to poor design decisions across the board.

For example: Playing as a zombie you first enter the level as an invisible, noncorporeal "ghost" that can set up in a strategic location where you are not seen. However in L4D1 during the final level of a campaign, you lose the ability to position yourself strategically, instead spawning in at a semi-randomized spot. Since Special Infected / Zombie players can often be killed with 1 or 2 bullets, this is incredibly unfair and frustrating. In L4D2 this was finally fixed, after approximately 2 years of complaining, however now the game enforces that you spawn a minimum of 100 meters away from the survivors. However this minimum distance means that Survivor players can essentially prevent the other team from attacking by positioning themselves in a spread out formation.