Friday, July 28, 2006

Bursty Indie Sales Cycles

I had a delightful lunch today with Amanda F. and her handsome, watch loving friend. She is the driving force behind the new indie RPG Aveyond and I’m very much looking forward to seeing what she does next. Most of my contacts with the game industry go back to the old PC shareware glory days, so it is quite enjoyable to connect with one of the rising stars of the new generation of entrepreneurial game developers.

Out of the many topics we meandered through, one jewel was the bursty nature of shareware game traffic. She’s been noticing a trend. Whenever her game hits a new portal, there is a rise in traffic across all portals that the game is featured, even her website. Portals where her game has long fallen off the chart suddenly start featuring her title again.

There are two potential reasons here:
  • Repeat impressions are needed before customers take action.
  • The downloadable market is highly fragmented.

Repeat impressions matter
People don’t look at a game and think “hey, I’ll try it out.” The first time, they become aware of the title, they might be about to wash their laundry or perhaps they are at work. Maybe they aren't in the mood to check out games. (Shocking!) The moment passes and the title that has consumed a year or more of your life passes out of their heads without a second thought.

Getting people to download your game is a lot like playing one of those maddening quarter games at the arcade. The machines taunt you with dozens of quarter balancing precariously on the edge of a small ledge. All you have to do is place in a single quarter and you’ll push an avalanche of coins over the edge.

But imagine that you start with an empty machine and each quarter is actually a mention of your game. You need to build up quite a few impressions of your game within a potential buyer’s head before the cascade of impressions overflows into action.

When a title hits a new portal, there is a buzz of word of mouth around it. This leads to lots of people getting fresh impressions of your title. Only a few are saturated with enough of your message “Hey, this is a cool game” to actually take the extraordinary effort to search the internet and download it. This leads to more word of mouth and more downloads. Thus the single media event leads to a burst of sales across multiple distribution sources.

Market fragmentation
The fact that the downloadable and casual games market is fragmented isn’t really news, but it too informs your sales patterns. Think of the new shareware market being composed of dozens (if not hundreds) of population pockets. Each group might be built around a single portal or a special interest group.

They don’t talk to one another much, nor do they read common news sources. Many don’t consider themselves mainstream gamers. I like to think of them as the oil shale of the gaming market: A bit difficult to reach in larger numbers, but still highly valuable customers if you can figure out the techniques.

The result is that long term promotion will often have incremental payoff even with products that a no longer ‘hot’. There will almost always exists large populations of players that will have never heard of your game. Don’t be surprised if you end up getting letters years after your initial launch that exclaim “I had no idea that this [insert superlative] game existed!”

Often someone who just heard about your game may introduce it to new markets. Within a short period of time, the number of people who become aware of your game can increase dramatically. This also contributes to and magnifies the bursty nature of sales.

You must pop little markets one at a time over a long period of time before the total number of customers that might buy your game is tapped out. Indie games are in many ways closer to evergreen products than your typical launch and dispose commercial titles. Think about it. Bejeweled is still selling to this day and it is doubtful that the majority of those customers are repeat buyers.

Marketing is a long term effort
There is the dark side to all this as well. If your game doesn’t trigger a big enough burst of word of mouth, you may see a small spike that fades away rapidly. Quite likely your awareness raising event isn’t large enough to ignite a chain reaction across all the sparsely connected social nodes. Alternatively your game isn’t good enough to inspire strong word of mouth. Or maybe you are popping smaller markets and not reaching the bigger ones.

I think of the system as the following (completely unscientific) equation:
  • [# of promotional events]
  • * [Average reach of promotional events]
  • * [Word-of-mouth worthiness of your title]
  • * [Average number of existing impressions]
  • * [Number of new markets that you breach]
  • * [Average size of each mini-market]
  • - Percentage of market already reached.
  • - Percentage of people who just don't give a damn.
  • = Magnitude of each PR burst.
You’ll likely have to promote your game for longer than expected. Don’t give up on an older title just because it is no longer the latest thing. Re-releases, targeting radically different audiences with an existing products, as well as shameless and consistent broad-based self promotion are all valid and useful techniques for getting your games out in the public eye.

Here is to a long and bursty sales cycle,

Links and such

A description of that darned quarter game

Monday, July 24, 2006

Games are designer food for infovores

I happened across this article in New Scientist discussing how the brain processes information. Research by Irving Biederman of Universtiy of Southern California in University Park and Edward Vessel of New York University provides some indirect scientific backing for many of the concepts I’ve discussed here such as:
  • Games as drugs
  • Burnout
  • The pleasure of groking that Raph Koster has discussed so eloquently.
They also claim to have coined the term “infovore” (which already had 86,800 hits on Google :-) I'm a big believer that there exists a strong foundation of neuroscience underlying the highly predictable behavior of people playing games.

Attempting to interpret information gives us a high
Scientists are beginning to understand the exact mechanisms in the brain that encourage the release of pleasure when consuming stimuli.

They claim that the neural pathways through which we learn about the world tap into the same pleasure networks in the brain as are activated by drugs like heroin. […] These are the areas that become active when the brain is trying to interpret the information it is receiving, whether that is an image of an object, or words on a page, or the song of a bird. Biederman and Vessel suggest that when this happens, the endorphins that stimulate mu-opioid receptors are released, causing a feeling of pleasure.

How to increase the high
We can improve the impact of our gaming systems by using information-based rewards that are highly pertinent to our audience.

What’s more, because the number of mu-opioid receptors increases the further along the neural processing pathway, information that triggers the most memories and conveys the most meaning to a person causes the greatest pleasure response. It is this bonus that compels people to browse for new information.

Burnout is briefly discussed. I’d like to see a lot more on this particular topic.

Does this effect ever wear thin? Yes, with repetition. Reading a book for the second time is less stimulating than reading it for the first time. “

Delaying burnout
The article also describes how delayed comprehension entices people to keep coming back to the same information source.

Beiderman and Vessel say that endorphins are released at the ‘click’ of comprehension, and that until the penny drops people are happy to return to a subject. Children take longer to “click” than adults – which explains their enthusiasm for hearing the same bedtime story night after night.

Quite fascinating, really.


Sunday, July 23, 2006

Ze Story Snobs

There exists a powerful group within the gaming world that actively seeks to stamp out innovation unless it falls along the prescribed lines of their rigid and conservative doctrine. Who are these people? They are every developer, gamer and publisher who promotes the ideal that a good game must have a story.

Raised on fine stories from the golden years of novel writing and movie production, story snobs see games as just another opportunity to tell great tales.

Poor David Jaffe fell victim to the bitter wrath of the snobs recently when he talked about how developing story based games isn’t all that exciting. It was an honest comment that makes a lot of sense if you’ve ever experienced the joy of tweaking a surprisingly interesting interactive system versus the slog of polishing a series of plot points.

It is really very simple. Not all games need stories. Treat story as one of many available marvelous ingredients that can improve your game, not as a necessity.

The logic of the Story Snobs
  1. I like games with stories!
  2. There aren’t as many great games with stories as there are books and movies with great stories.
  3. It is therefore the fault of [the developer, publisher, etc] because they are not filling my needs.
  4. As an advocate, I must passionately protect and promote any game with a story as the ideal.
  5. Anyone who suggests games without stories are reasonable should be crushed. After all, it is a zero sum game here. Any resources spent on promoting non-game stories are resources that could have been spent on 10 more dialog trees.
The root of this unfortunate attitude is a classic tale of old media infiltrating and co-opting a new media. There are three players:
  • The fanboys
  • The movie and book industry wannabes
  • The publishers

The Fanboys
There exist millions of fan boys who had great experiences with old adventure titles and Japanese-style RPGs such as Final Fantasy. These story rich titles were some of the first cross over genres that encouraged people not typically interested in games to pick them up and try them out. If you are a conservative media consumer used to movies, Final Fantasy is an easy dish to consume. You have to watch a few cut scenes, play a little bit of game and watch a few more cut scenes.

However, many of these new game players never moved onto new genres. Just like a good number of Brain Training new customers never try racing games, there are millions who started playing games with the adventure game genre and stopped when that market faltered. There are millions that to this day still play mostly Japanese-style RPGs.

For this demographic, the artificially sweetened formula of “Lots of plot with a dash of interactive bits” defines their total vision of gaming. As conservative media consumers, when game falls outside their nearly religious preferences they don’t merely accept and forgive. Instead, they are inclined to drag it behind their truck through the proverbial forums of Texas. “A game without a story? Impossible.”

Despite the copious evidence to the contrary.

The movie and book industry wannabes
I had a great conversation with an animator at GDC. He’s an industry veteran and works primarily on cut scenes. He confided to me that his true dream was to work on CG for movies. He read all the movie trade magazines and avidly sucked up their tips and techniques. He was in the game business because it was kind of similar and he could get a job there.

I’ve had roughly this same conversation with a remarkable number of developers. The game industry is filled with writers who want to author the next great novel, designers who want to direct the next great movie and artists who would be perfectly happy doing character design for a Saturday morning cartoon. Even if they aren’t actively trying to use the game industry as a stepping stone, many of their core values are informed by older, existing media such as movies or novels.

These cultural transfers from big established media industries have a huge impact on the type of games that are made. First, their general grasp of how interactive systems are built is quite weak. They couldn’t design a set of valid game mechanics if they tried. More importantly the passion for interactivity amongst many of the developers in the game industry is unexpectedly low. When you talk about making a sexy Blizzard-style rendered intro, eyes light up with respect and admiration. This is their dream. When you talk about emergent gameplay in a title like GTA, you’ll get blank stares. It just isn’t their passion.

If you have the skills to make movies, everything looks like a movie. There are a thousand decisions made during game development that are the creative choices of the developers involved. If your labor force is trained to build and steeped in the culture, and aesthetic of linear media, guess what most games will end up looking like? That’s right. Linear media with chunk of half assed or cloned interactivity thrown in for good measure.

I got a chance to read a game design document for a now published title. It read like a movie script. Except they had little production notes like “And now the character fights a red monster”. Interactivity in games should be more than just a production note.

But it never will be when large portions of our industry’s workforce worships the values of linear media over the unique charms of interactive gaming.

I can’t blame the business folks too much. They have their creative people telling them that stories are critical. They got violently passionate customers telling them that stories are the most important thing ever. So they do what sheeple do and green light mostly story-based projects.

The vast majority of the budgets in modern games goes towards art, video, dialog and other plot related expenses. The development teams are further stocked with Hollywood refuse, which only increases their story-centric biases. Game mechanics work is generally given less development time, resources or room for experimentation.

Since the production risk of story-based games is lower, publisher tend to green light them more often. The developers don’t know how to replicate the complex playground games that do become hits. The market ends up being flooded with dozens of story-based games and only a few games that focus on interactivity as the primary driver of value. So we train more players to expect story-based games and we train or import more developers that know how to only make story-based games.

The industry becomes more and more weighted towards producing games with stories. You end up with a feedback cycle that reinforces the required presence of story elements in most games. If everyone wants story, how can it be wrong?

As various folks have commented, Tetris would never be published today. The current requirement that most games must have stories is a filter that prevents the creation and publishing of what are potentially the crown jewels of the gaming industry.

There are lots of great games that don’t require a story. Focusing our effort on only creating games with story substantially limits our creative exploration of the media and limits the types of games that we, as game developers, are encouraged to create.

Stories are not required to make great games.
Before you think I’m a story hater, let me disabuse you of the notion. I like stories. I’m playing a darling little RPG right now called Aveyond that is quite plot heavy. Delightful stuff that simply would not work without the inclusion of a story.

Even as a story lover, however, the existence of the story fiends infiltrating every level of gaming irks me. They assume that stories are always a good thing. People are not thinking critically about whether or not their game needs story elements.

It is perfectly possible to have a great game whose plot elements fit on a postcard. Populous, Mario 64, Quake, Lumines, Bomberman, Guitar Hero, Counterstrike and hundreds of other titles succeed wildly as great gaming titles and yet all of them lack story beyond a rough setting. They don’t feed the player periodic plot points that extend a narrative. They don’t have characters with extensive histories that evolve and grow emotionally through a series of descriptive cut scenes. They don’t have fixed events that are described by a godly author as a way of informing the player about actions beyond the capabilities of the gaming system to simulate. And they rock none the less.

In fact, the one thing that prevents the game industry from turning into Hollywood with occasional button pushing to advance the plot is the fact that a lot of people purchase certain hits that shockingly have little evidence of tradition plot. Sports game and racing games consistently make a profit. Nintendogs and Brain Training came out of the blue and rocked Japan. Tetris made the Gameboy a success. All these smash hits have no plot and lots of interactivity.

So there is obviously a more complex tale to be told here. There exists a wide swath of games that can be successful without having a story. Just as there also exists a wide number of games that can benefit from having a story.

When players and developers simply assume that their games need stories because they have been blinded by their subconscious cultural biases, they fail to dig into the guts of why stories matter to games. When you say “Wouldn’t it be nice to have another cut scene because I like cut scenes,” you typically aren’t asking the hard question “What does this cut scene actually bring to the gaming experience as a whole?”

Story is a game design ingredient, not an end in and of itself.
Story has a purpose in game development. It is a ingredient. It has little inherent value by itself. Its primary value is how it contributes to the entire player experience. You are selling a game, not a movie or a novel. You need to design the whole game as a complete experience.

To use a bizarre analogy, making games is a lot like cooking. You may really like bleu cheese. I do. I once found a fabulous recipe for bleu cheese lasagna. The recipe called for a few crumbles of bleu cheese, but the store only sold large hunks. I thought to myself “I like bleu cheese a lot. Why waste all this cheese…I’ll just throw it all in!”


All the bleu cheese went in along with some expensive spices and other goodies. The result was a giant slab of goo that tasted intensely of bleu cheese. You couldn’t taste the spices, the sauce or the noodle. I ate it for two weeks straight and never ate bleu cheese again for months. I would have been better off just nibbling on the chunk of bleu cheese.

I added something in that I liked by itself, but I didn’t have a clue about how it would interact with or benefit the other elements in my dish. The same goes for gaming elements like story. What do they add to the game? If you end up with a game that is barely different from a movie, why not just make the movie in the first place?

What is a story to a game?
Let’s take a look at the role story plays in game development.

First off, it is worth defining story. Story is a series of linear narrative elements in the fashion of novels and movies from ages past. This is a very traditional definition that I’m confident does a disservice to many of the wacky interactive fiction attempts being concocted by mad geniuses around the world. It also happens to be the one that is most descriptive of the use of story in modern video games.

In most games story elements are used as rewards for player actions. The player does something and they get a little dose of plot. Typically plot points fall into one of three categories.
  • Enabling reward: These rewards help the player advance through the game further. Examples include the conversations in Half Life 4 that let you know that the main generator is down and needs a fadangle to fix it.

  • Red herring reward: These are rewards that the player instinctually pays attention to as potentially important, but in reality they are just tossed in there to help build a fantasy world. The player, not being able to distinguish between what clues are pertinent to the game world, laps the red herrings up and experiences the same sort of pleasure they would gain from an Enabling reward. Examples include descriptions of a Dark Past Foozle that once caused a huge cataclysm. It never affects the actual game, but players latch onto it and try to make sense of it none the less.

  • Visceral reward: These rewards trick our sensory system into thinking something interesting is happening. Example include big bloody fight scenes, spooky scenes that cause us to think we are in immediate danger even though we are actually sitting in a comfy chair in a posh apartment on the west side.

The model we are using here assumes that gamers are constantly trying to grok the gaming world in order to interact with it in a more meaningful manner. The primary bursts of pleasure come from activation of learning systems in the brain. There are secondary burst of sensation that come from false sensory input that activates various fight or flight mechanisms. It is a simple model, but it generally works and is a far better starting place than designing by feel alone.

When should story be used in a game?
So a story element is just a reward. It isn’t the only type of reward. It is one of many types of rewards. You could put in a cut scene when a big boss creature is destroyed, or you could let the player discover a new sword token that enables them to chop down the vines that have been blocking progress through the earlier jungle levels. Both might cost the same amount of development time and both are valid rewards that make the player feel great.

Instead of asking “how should I implement story in this game?” instead ask the question “What type of rewards best fit the game experience?”

Story-based rewards have several very distinctive characteristics that can influence your decision.
  • Triggers for specific types of emotions: The biggest benefit of story-based rewards is that you can use them to trigger social emotions such as sadness, humor or sympathy. These are typically difficult to trigger using algorithmic rewards, but are relatively easy to create using common narrative techniques and patterns.

  • Low initial production cost: With a line of text, you can hint at complex system that you never need to build. For example: “The tattered scroll describes the ancient history of Yendor where giant lavender airwhales ruled the skies” hints at a tantalizing other world that the game developer will never need to build. The cost? A few minutes of writing in a commonly available word processor. In general, a simple plot point can be created and polished at a much lower cost than what it takes to create an interactive reward.

  • Rapidly escalating costs as realism increases: As you attempt to increase visceral aspect of your story, production costs increase dramatically. Realism costs money in the form of expensive tools, talent and time.

  • Low execution risk: The risk of a story-based reward failing to be completed is very low. The production techniques for text, images, sounds and video are well understood and highly reproducible. If new technology is kept to a minimum, the use of story-based rewards is highly unlikely to delay the shipment of your game.

  • High burn out: Most story rewards have very low variability. When you see them once, you’ve sucked out 99% of their value. When you see them again, players get much less buzz. Repeated often enough and they become downright irritating. Imagine if you were forced to watch the main intro animation when you start up a game. It rapidly loses its appeal.

  • Limited economies of scale: With high burn out come very limited economies of scale. Every time you want a new reward, you need to custom craft a new one. The cost of generating new reward increases linearly with the number of rewards. More algorithmic systems, on the other hand, tend to have a higher initial cost but can be reused over and over again. Imagine having to come up with a unique and enticing story element every time the player killed a monster. It is much cheaper in the long run to simply give the player a few points or a health pack. Also note that the cost of a small increase in realism is multiplied across all the story rewards in your game. This gets expensive quicly.

  • Changes are expensive: Exploring variations on story elements is expensive. Often the plot points need to be rebuilt from scratch. If you are dealing with text, this isn’t so bad. If you are dealing with $50,000 cutscenes, it can be quite painful. Contrast this to more algorithmic system were changing drop percentages on rare items may be as simple as tweaking a single number and seeing what happens.

There are some folks out there who claim that stories work poorly with games. This has some root of truth. If you focus only on the quality of the story, you’ll find that your gameplay elements will appear to constantly interrupt and slow down the flow of the story compared to say your favorite movie. It can be difficult to built dramatic tension in a typical manner when the player is constantly jumping about, pressing buttons and performing other mechanical actions.

However, story can still be used as a vital element to the game. It can add an emotional richness to the reward system that is difficult and expensive to achieve using algorithmic techniques. Story must always serve the greater good of the game.

To return to our cooking metaphor, I find story to be much like a fine wine. When you pour a glass of gloriously rich merlot, you’ll discover all sorts of delicate nuances that are simply impossible to find anywhere else. Yet, that same glass of wine can also be used to cook some wonderful dishes. A nice lobster bisque wouldn’t be complete without a dash of wine to accent the flavor. Stories in games are really the equivalent of cooking wine, an essential and useful ingredient for many popular genres.

On the other hand, there are lots great dishes that you can cook without using any wine at all, just like there are some great games you can build that don’t use story. If game design is anything like cooking, there is an entire universe of game designs that work perfectly well without any story elements.

If you are yourself a story snob (I was for many years), you need to ask yourself “Am I in this business primarily to build a great game or am I in it to craft a great story?”

Pick your passion. If you really want to make movies, go for it. Move to LA, buy a video camera and get started! For those who choose to remain in the game industry, we’ve got a unique and wonderful medium that deserves to be explored and expanded as a powerful expressive force in its own right. Learn from old school linear media, but never be bound by its constraints. Use it as one ingredient in your dish only if you want a dash of story flavor.

Don’t be a story snob and assume blindly that your game needs a story. Players buy games for the total experience and you should choose the appropriate reward system that best fits the experience that you are attempting to craft.

Take care


How the story fiends filter great games
“I believe if Tetris were presented today, here is what the producer would be told: Go back give me more levels give me better graphics give me cinematics and you re probably going to need a movie license to sell that idea to the public. The producer would go away dejected. Today, Tetris might never be made.”
Satoru Iwata, GDC 2006 Keynote

Burn out: When you create story elements as rewards, you experience the same reward over and over again as you play test the title. This leads to burn out on those rewards.

“And the thing is, once you have the IDEA, your fun- as a designer- is really over. If you are working in the single player action-adventure genre, and are fortunate enough to be working with a team that can execute the crap out of what you think is an amazing idea, you don’t get much out of actually seeing your idea executed. You get a little, sure. You get the little tinglies and such. It’s a neat moment to see your idea brought to life. But you already saw the idea, already experienced the amazing moment...but it was in your head months ago. Now it’s just a slog to execute the damn thing so OTHERS- the PLAYERS- can enjoy what you’ve already finished enjoying.”

Lobster bisque with wine

Saturday, July 15, 2006

Playground game design as a sustainable competitive advantage

There exists a perplexing multi-billion dollar mystery sitting in plain sight of everyone who works in the game industry. Some of the top franchises such as GTA or Sims do not experience the same competitive pressures as do other titles in popular genres. Also unexpected is that the companies that originally innovated with the creation of a new genre end up dominating. Something beyond your typical branding and IP ownership acts as a barrier to entry for new companies looking to cash in on popular new genres.

The urge to exploit these obviously successful genres seems impossible to ignore.
  • Open ended dollhouse genre opportunity: The Sims has sold over a billion USD of product (24 million copies including expansions) since its introduction.

    The Competition: A mere half dozen failed attempts such as Singles or Playboy Mansion.

  • Open ended driving genre opportunity: The last three versions of GTA have sold over 38.44 million copies, placing them amongst the most successful franchises of all time.

    The Competition
    : Around a dozen urban themed driving titles, most of which miss the entire point of the GTA design by removing the open ended gameplay and implementing linear levels. Most started development only in the last couple of years. There is an intriguing lag of years after the signals of market success for the genre hit the development community.
Compare this pattern to the RTS rush that happened a few years after Dune 2 was released. At one point, I remember counting over 60 different announced RTS titles. The same genre rush has happened with FPS, Match-3, Vertically Scrolling Shooters, Graphic Adventures, and most new successful genres throughout the history of the game development.

What prevents the typical genre gold rush from flooding the market with profit sapping competitors to the Sims or GTA?

Playground games
One of the few common elements that titles like GTA, Sims, or Oblivion sport is open ended gameplay where the player explores a large, complex playground and meanders about from non-linear adventure to non-linear adventure. The concept goes back to a couple of decades to early titles such as Elite, Nethack and Paradroid. Each of these titles focuses strongly on dynamic emergent gameplay over heavily scripted lineary experiences.

Such playground games are bloody difficult to build. They balance complex emergent gameplay systems that, by definition and design, lob crazy surprises at the development team all throughout development. With enhanced player freedom, there arises an inevitable stream of exploits, bugs and worst of all unbalanced boring game play.

The flip side is that playground offer unexpected competitive insulation from the typical cloning efforts of other development teams. A good playground game has a magic mix of gameplay that is both difficult to get right in first place and difficult to duplicate. The recent spate of high budget attempts on GTA may replicate the urban setting and adds better graphics and more explicit violence. But none of them capture the gameplay magic.

A unique sustainable competitive advantage
This unexpected quirk of playground games is a potential pot of gold. It turns out that any difficult to replicate process that yields a valuable product can be used to help make us more money. It opens up the possibility of milking new games longer without the threat of competition.

Any venture capitalist and most entrepreneurs are familiar with the concept of a ‘unique, sustainable competitive advantage.” Traditional economics teaches us that most highly profitable new products are converted into low profit commodities through competitive pressure.
  • A company releases a new product that meets a hitherto underserved market need.
  • The first mover, recognizing that they offer a unique product, charges a premium above and beyond their production costs. In situations where demand outstrips supply, the customers happily pay what my economics professor termed “greedy pig profits.”
  • Other companies recognize the sweet scent of profits so they duplicate the original product, add some of their own mild innovations, and typical offers a competing product at a lower price or with a bigger marketing budget.
  • The original company responds in turn, which naturally costs money. There is a bit less profit margin for everyone at this point.
  • After several cycles of increasing competition, you end up with a mature market in which each company is spending almost all their profits on marketing, distribution, and product improvements. The era of the greedy pig profit is at an end.
There are quite a few other potential endings to this classic tale. The one you most typically see in the game industry is one where the original company is often bankrupt, purchased or marginalized. A few bigger companies that can leverage economies of scale or market dominance take control of the new market and release a steady stream of moderately profitable products. They erect substantial barriers to entry in terms of licenses, strong branding, or high costs of entry.

Anyone looking at a new business (and most new game projects are new business projects) understands these dynamics. They know that a great new innovative product you make today will likely be cloned, rebranded and controlled by the competition tomorrow.

So as educated business owners, we must ask the dreaded question “What is our unique sustainable competitive advantage?” In essence, we need to figure out how we are going to break the inevitable capitalist cycle that drives those lovely greedy pig profits out of our business. What barriers to entry can you put into place now that buy you time to rake in more money for a longer period of time?

I suggest that complex emergent game mechanics are one mechanism that acts as a sustainable competitive advantage for new games. We've all been trained to appreciate typical stalwart strategies such as brand building and licensing. A title's game design, oddly enough, is rarely seen as a key business decision. It is one of those artistic details where ephemeral 'quality' matters, but has little impact on the bottom line. What if picking the right class of game mechanics could result in the creation of defensible franchise that profited your company with profits for decades?

The naive look at the roll of game design in business
Many game developers imagine that inventing a great new game is enough. However, with most modern game designs, the core game mechanics can be duplicated relatively easily. There are fewer moving parts and their impact on the overall gameplay is clearly defined. Once there is a proven working example of the game design, the barriers to entry are quite low. Within a couple of years, consumers witness dozens of clones clogging the market. This is known as the "genre gold rush."

In a market where prototypes are often more effective than design documents, releasing a typical game is like hand delivering a blueprint of your entire company’s formula for success to the doorsteps of your competitors.
  • Blizzard stole the RTS crown from Westwood (who is now dead as an independent company)
  • Everquest stole the US MMORPG crown from the UO team (and Blizzard in turn stole the crown again with WoW.) The original UO team is now marginalized.
  • The various Sly, Ratchet and Jax titles now own the 3D platformer genre. Nintendo, with their original Mario 64 franchise, is rarely seen. This ‘innovate and move on in the face of competition’ is arguably a successful business strategy for Nintendo, but it still leaves a remarkable amount of profit on the table for competitors.
Wouldn’t it be nice to cut this cycle off at the knees? As a game developer, you ideally want to create a new genre and milk it for the next ten years without the threat of outside competition.

How emergent gameplay results in a sustainable competitive advantage?
This is exactly what has happened with GTA, the Sims, Oblivion and small number of other titles. By investing in a complex sandbox gameplay system, they’ve created a strong barrier to entry. Typical game development companies that thrive on cloning the works of others can’t get their heads around the formula. The reasons why the barriers are so strong is a fascinating lesson in the failings of traditional game design techniques.

Every post mortem that I’ve seen for titles with strong emergent gameplay in an open ended playground like environment discusses the long, convoluted development process. The Sims languished for years as Will Wrights personal project and originally was going to be an architecture simulator. GTA at one point didn’t have cars and involved killing zombies. NetHack has been in development for decades with the community adding quirky mechanics in a haphazard fashion to the system.

Some common themes include
  • Use of iterative design to polish and improve as the design goes
  • Openess to big design changes well after the typical preproduction stage.
  • Layering multiple unique gameplay systems on top of one another to form a complex choice space for the player to explore
  • Long development cycles that encourage the ‘ripening’ of new game design concepts.
These successful best practices also introduce the following entry barriers to traditional ‘clone and polish’ companies:
  • Design risk: The existence of so many different systems interacting in intricate ways means that simply cloning the mechanics exactly rarely results the same emergent behavior. There is an element of the butterfly effect involved. A minor difference in the city generation algorithm combined with a small change in car handling physics and certain types of jumps and stunts become highly unlikely. So the chances of your title containing identical moments of ‘fun’ as the original are highly unlikely.
  • Process and Team barriers: Initial versions of playground games are almost always developed using highly iterative, prototype focused processes. Many teams do a small amount of prototyping early on their development, but they are not set up or trained to manage an iterative process over the entire length of the project. Teams trained on “Plan and then Execute on a Production Line” often fail to have the personalities or experience to successful carry out an extended “Prototype, Riff, Critique, Expand” process.
  • Risk adverse publishers: Iterative design is messy and highly frightening to most risk adverse publishers. The metrics used in the industry today are such that when a project with highly emergent gameplay is meandering towards a destination, it is very likely to be cut. No one likes hearing about major game changes months before release, but that is how these emergent game systems evolve. “Oh, btw. We added a new driving system.” The publisher will inevitably ask “Why can’t you add three more levels exactly like the existing ones, render a great box shot and just ship it?” If the standard formula doesn’t work, then the game must be a bad egg.
It is a rather vicious cycle. Companies that look at cloning a popular genre tend to use risk adverse best practices like data driven development. Their goal is to reduce risk through upfront planning and detailed execution of the plan. These techniques are the opposite of what you need to build a new playground game. In iterative development, you are constantly learning from the emerging gameplay as it evolves on a daily basis. You tweak, evolve, build a little more until you have an enjoyable organic experience.

However, when the team judges the progress of a game based on how well it sticks to the production plan, they are often blind to finding and nurturing the magic moments that make a playground game function. The title almost inevitably suffers from an unexpected case of ‘crappiness.’ With the release at risk, a crack down by upper management ensues, often involving more detailed planning. These are ‘sane, reasonable’ best practices smother the processes necessary to create and balance a playground game. By trying to reduce risk, they actually increase their risk of market failure.

Benefits entry barriers
Not having competition from the clone-meisters of the world has strong benefits for a company.
  • Time at the top: Obviously having a successful title that people love allows you to release a sequel. Without competition, that sequel is very likely to be a genre king. All the press, word of mouth and pent up demand from previous players translates into a much larger swell of people purchasing the newest top title in the genre. If you have the only option, you reap the benefits. Having a third or fourth uncontested genre king in a row results in major cash in your bank account.
  • Brand: Without competition, you have lots of time to establish your brand. The first Elder Scrolls title was a mild PC success, but nothing to write home about. Over the past decade, they’ve managed to build a substantial brand following that will be quite hard for others to erode with competitive products. If Oblivion II and an exact clone of Oblivion II were released simultaneously, Oblivion would blow away the competitor in terms of sales. At this point, the brand has established market power.
  • Team empowerment: Since the team dynamics are so important to the success of the franchise, playground teams generally are given great creative freedom by the publishers. Such teams are literally golden gooses that make or break the bottom line of those that are lucky enough to have them in their development stable. Designing what you want, when you want is not such a bad thing.
Problems with creating playground games
All this sounds great, but there are some more big reasons why more people aren’t making playground games. The obvious ones involve the same barriers to entry the stymie the clone makers. However, even when you are doing iterative development there is a high risk of failure. I’m reminded of Frontier, one of the sequels to the great playground title Elite. This was a game that had a lot of things going for it: brand, original developers, and proven concept. But it ended up suffering horribly in the market for a variety of reasons.
  • Bugs: Complex systems introduce crazy amounts of bugs. Stability is often an issue unless you architect the system to deal with rapid design changes.
  • Indeterminate development cycles: It is often difficult to know when to stop. There is always something new to add so development could literally go on for years. Reducing scope and remaining open to change is a hard balancing act for many teams.
  • Team dynamics: More than any other type of game, who is on the team matters. If your team doesn’t gel, the open opinion rich work environment will result in destructive conflict, not creativity.
The result is that many playground games fail. Very few have the resources or the right people on board to wade through the difficult times that inevitably crop up. Each one is a unique creative adventure that is just as likely to be stoned by an unexpected cockatrice as it to produce the Amulet of Yendor.

We live in a turbulent industry where developers move from team to team and project to project at an alarming rate. Stability lasts as long as your teams signed titles development and then all bets are off. As developers, we generally do not understand how to carve out a sustainable niche in the marketplace and build a lasting company around that niche.

Yet, companies are doing just this. They’ve built their own sustainable niches by inventing new genres that are easy to defend against competition. Bethesda, after many years of hard work, has a franchise that will feed them for many years to come. The old DMA, albeit much changed from those early days, still has years of GTA left in them. Will Wright, a quirky guy building quirky games, is in the position to infect much of EA with some of his iterative prototyping methodologies. He has shown the world’s biggest publish how to build a competitive advantage that was based off game design, not licensing and branding.

If there is one take away from this article, it is that building playground games is a sound business strategy. Building playground games is not easy and very often, you will fail. You need to build your teams around iterative design. You need the right people on board and the long runway necessary to ripen your title. Spending a decade honing a single concept is not uncommon. However, if you get everything right, you will have created a long lived, stable source of greedy pig profits.

There is the bigger picture here as well. If you pull it off, you've generated stable employment for your team. They don’t have to move across the country every couple of years to take up a new title. They don’t have to pull their kids out of school or take a job that isn’t what they love in order to avoid selling the house. Your team gains a small amount of freedom from the rat race.

This is a good goal and there are admittedly lots of ways to reach it. Playground games are one more sustainable competitive advantage to keep in your quiver. They can help you do what you love and still have a good life doing it.

Take care

Appendix A: Lessons from Grand Theft Auto
Here are some quick notes I jotted down after looking through the post mortem that Edge did on GTA. Here are a few lessons that can be gleaned from Grand Theft Auto. I've witnessed many of these personally and also seen the same patterns pop up in other playground game post mortems.
  • Sticking with the playground concept
  • Slack: Time to experiment
  • Permission to experiment
  • Collaborative
  • Iterative game design
  • Willingness to trash the setting
Sticking with the playground concept
“They wanted something more convincing, more immersive. ‘Living’ environments seemed to be the answer”

The big concept driving GTA was to build a ‘living city’. The further they went down the path of development, the more this playground metaphor became the driving theme of the title. Instead of reducing the experience to a series of canned levels, they opened the world up to an entire city and made the mission optional.

If you are making a playground game, don’t give up on the concept because it is difficult and risky. You toss your competitive advantage out the door the moment someone says “Maybe we should split the game up into easy-to-manage linear levels.”

Slack time
“But, fortunately, Dailly had been working on another, unrelated idea.”

There are two types of risk. The first is execution risk, which is the risk that a project will not be completed. The second is design risk, which is the risk that you’ll complete the wrong project. Most companies focus on reducing execution risk by having tight schedules and detailed planning. Unfortunately, this is often like driving a fast car in the wrong direction.

Design risk requires the time to generate new, potentially better ideas. If you are driving your team 80 hours a week to check off the 2045 little items on the Gant chart, they will never have the time to come up with something better. It just won’t happen.

They need slack time to play and come up with their own ideas.

Permission to experiment
“If we’d stuck to the original design, GTA would have flopped.”

The other aspect of design risk is an openness to incorporating new ideas into the product.

If you don’t have the ability add new stuff to the project, the team’s creative spring will dry up and the design risk for the product will shoot through the roof. In GTA’s situation, not only was the team onboard, but the producer was comfortable with the experimentation as well. The team was given another 12 months to get the title right.

Collaborative design
“The one thing that everyone agrees on is that they didn’t make Grand Theft Auto, but that’s not strictly fair: the other thing that everyone agrees on is that everybody made Grand Theft Auto.”

Often, a purely production oriented studio will rely on a command and control style management structure with strong role boundaries in place to prevent any disruptive discourse. Suggestions are required to go through proper channels with titles such as ‘Lead Designer’ or ‘Producer’ acting as strong gatekeepers that prevent most ideas from making it into the project. By ensuring that individuals own decisions, you guarantee that there is someone to blame if things get out of hand.

You end up with a carefully structured dance of petitioning the higher powers in order to ensure that a feature makes its way into the development pipeline. The nobility of the team that best plays the political games has the most say in the creative process. Most of the effort goes into navigating the system or complaining about the system instead of thinking up new ideas. In order to reduce distractions, many managers inadvertently implement rules that disable the creative engine of the team.

An alternative is the use of small teams that openly share ideas and are jointly responsible for making sound decisions. The team takes ideas from any source and builds upon them to reach a superior solution. To keep the creative engine burning hotly, everyone must be able to contribute to the design.

“I sat through heated design meetings, which resulted in tears. Screaming, punches and arguments were common.”

When you are practicing collaborative game design, a passionate team must figure out processes for embracing and dealing with substantial ambiguity. Successful team strategies for dealing with conflict will often result in confrontations and arguments. Ideally, you get passionate conflict that causes people to think through, question and defend their ideas.

There is no one answer to dealing with this inevitable conflict. Each team ends up setting its own norms and the teams that gel successfully are another difficult-to-replicate competitive advantage.

Iterative game design
“One of the programmers came up with a routine that had pedestrians following each other. This led to the idea of a line of Krishnas following each other down the street and then, once we had all experimented with ploughing through them all in one go, the Gouranga bonus became an obvious addition.”

The core of the design process for GTA involved an iterative prototyping cycle
  • Try out an interesting idea
  • Ask “Are there any interesting situations that the player finds themselves in?”
  • Build reward system around the interesting situations. Expand promising systems to allow the player to experiment in those areas further.
Again, this is quite different from the typical design bible that drives most production oriented companies. The team explores, prunes and expands upon the gameplay space instead of following a rigid production schedule.

Willingness to trash the setting
“Many ideas and approaches were bandied about, some quite different to the final game”

Often games are pitched with a particular setting such as a “Vampire action game.” Changing the setting requires big changes to the design document, the plot and the art resources. By risk adverse publishers this is typically considered a Very Bad Thing.

One of the fascinating things that happens with iterative design is that the setting changes as you go. GTA switched settings from being a gang warfare title to a zombie game to a car jacking convict game. As you try out new mechanics, they map onto different settings. Since the focus is on polishing the game mechanics, setting takes a back seat. As your game mechanics evolve throughout the game, so does the setting. The results are often unexpected and quite delightful.

The Making of Grand Theft Auto

Playground worlds

Game sales

Definition of the term ‘genre’
I refer to game genre as a group of games having similar game mechanics. Unlike movies or books, two games in the same genre can have completely disparate settings but still share the same mechanics. For example, Starcraft and Warcraft both use RTS game mechanics and therefore in the same RTS genre, despite one having a fantasy setting and the other having a science fiction setting.

Tuesday, July 11, 2006

Lessons from the Land of Pixel Art

I had a lovely visit to a pixel art website recently. Wow. What an amazing community.

Back in the day, the artwork I did was, if not cutting edge, at least quite close to state-of-the-art. Game artists who drew with pixels were pushing the technological boundaries of the time. They dithered because it was a clever technique to drag the sexiest look out of limited hardware. They mapped palette indexs to assembly coded color shifters. Color cycling was the equivalent of Spore's high tech procedural animation for the day. Give an artist a pot of mud and they'll make something beautiful. Limited palettes and square pixels were the mud that we were handed. We simply made the best of it.

What we made wasn't intended to be special. We created mass market images for disposable games titles. Mere years after release and you would be hard pressed to even find hardware capable of playing the titles. Pixel artists of yore pumped out art for consumption by pimply faced 12-year olds. We did whatever was fastest, cheapest and still satisfied our customers.

Little did we know that the art we made would one day be put on a pedestal and hailed as a "style." We ended up making real art that touched people emotionally and evoked beautiful memories. There are websites filled with rants about the superiority of pixel art and the misery of today's rampant use of modern 3D graphics. Pixel art, with its accidental association with the honey and spice memories of youth, now ‘means something.’

Ironically, the innovative spirit of those early pixel pushers now lives in the crazy brains of 3D modelers and shader writers. The bizarre mix of technology and art that goes into building the subtle beauty of Halo 3 comes from the same place as the fellow who said “Ha, I’ve figure out how to use 16 colors instead of just 8!” The urge to make great and evocative images will always press against the latest and greatest technological boundaries.

And there will always be those that are left behind. Those who came to game late, after the great works have been done, defend the traditions of the past with great fervor. It has happened with commercial illustration, commercial painting, film photography, print making and laying out graphics with a razor and straight edge. The same thing is happening with pixel art.

In the best of worlds, there is enough cultural momentum that the style establishes itself as a fine art, to be preserved on its own merits instead of its value to the consumer ecosystem. Will pixel art make the transition from mainstream art for the common man into a fine art?
  • Something of it may survive as a subtle form of mosiac. There are likely more blocky pictures of Mario drawn in medium like tiles, post-its, polygons and vectors these days than are drawn in actual pixels.
  • A few bands of connoisseurs may survive on the fringes of the art world, clinging to ancient copies of Deluxe Paint and insisting that only pure, organic pixels and limited high fiber palettes be used. No index painting allowed. (It is simply evil)
  • The surreal images of spaceships, mushroom and isometric landscapes will almost certainly live on, reinterpreted in new media with a wink and nod.
But ultimately, mass market art moves on as new generations look for new thrills. The practical artist, those commercial money minded drones such as myself, moves with it.

Right now, shader-based 3D is the new pixel art. How much longer until it too becomes retro and in need of protection by delightfully passionate fine art snobs?

take care

Saturday, July 8, 2006

More free game graphics

Here is another set of free game tiles for a 2D Zelda-like RPG that I discovered lurking on my hard drive. These were created for a prototype title so there are only a few sets completed. These would make a great start if you are in need of basic graphics for your next great game.

Wilderness Tile Set

Interior Tile Set

Village Tile Set

For the time, these were rather high tech.
  • The shadow tiles had a total of three levels of transparency built in so that you could get a deliciously extravagant soft shadow. The plan was to have some sweet assembly code that shifted our 8-bit palette by different amounts for each indexed shadow color. Soft shadows weren’t even a buzz word at that point.
  • You could have objects on top of other objects. Instead of drawing a table tile with a candle drawn into it, the candle could be a separate object layered on top of the table.
  • They were drawn in Painter 2.0 using a pressure sensitive tablet. They were my first graphics drawn in Painter. I never went back to Deluxe Paint again. :-)
All of this is of course trivial these days. If you’ve got a 24-bit 2D graphics engine, life is grand. I still get warm fuzzy feelings thinking about our mad plans though.

You’ll need to chop these up into the various pieces and then knock out the transparent color. It is a bit like a jigsaw puzzle, but it follows the same basic pattern of my other tiles sets. You can use the test pictures posted on this page to see if you are fitting everything together correctly.

I’d imagine that if you are making a 2D game with any sort of wilderness or villages, these would be an ideal starting place and are a lot less painful than tracking down a competent artist. As always, these tiles are free to be used in whatever projects you desire. (I need to get a copyleft license available at some point.) If you do use the graphics, drop me a note. I love hearing about projects.

Download the files here (1.89MB)