There is no feature, no asset, no line of code that doesn’t have a cost. Everything is a series of trade-offs. For every extra feature that goes into the box, you should ask why and how much. If a new feature, let’s say you want to add pedestrians walking dogs–just for ambiance. Additionally, let’s say that everything else in your game is bipedal. To add the dog you will need the art assets for the dog, a custom animation rig for the dog, and special AI to handle the dog-walking. You may even need special code to handle the leash (bending and flexing). You’ll need work from a modeler, a rigger, an animator, and at least one engineer. With implementation and testing you’re looking at at least four weeks. A simplified budget of $10K per man-month (this includes salary and overhead) results in a cost of $40K. For a dog. On a leash. That has no impact on gameplay. Bear in mind, that this is also one man-month that is not used for core game-play features.

This may sound quite a bit like feature-creepism, but the key difference here is that these features are included in the plan from the beginning. There is always a big list of “wouldn’t-it-be-cool-if” features that are added to the schedule at the beginning of the project: dynamic weather systems, ambient AI, etc. Not only do these features not have any real impact on the game in question, but they are invariably cut due to time constraints. But they are only cut after a significant initial investment of time and resources.

If you spend 100% of your time working on 65% of the planned project then you only have 65% of a game. If you stick to only those features that are necessary you can spend 100% on nearly 100% of your game. Recent titles in the “of war” genre, “God of War” and “Gears of War” are both excellent examples of features focused only on active elements of the game. You can see this reflected in both reviews and sales.

Recently the talk of episodic games has started to bubble to the surface again. The same ideas of episodic gameplay have been around since at least the mid-nineties. The difference now is that the digital streams to deliver the content are now widely available and a model for in-game advertising has begun to ferment.

It is not reasonable to compare episodic games with films, which are neither episodic not delivered, short of PPV and NetFlix, to the consumers home. Going to see a movie is still primarily a group experience. Looking at television may just provide us with some clues as to how to create and deliver episodic games.

First, highly-paid TV execs have already done the core research for us. Serial television, with the exception of reality shows, are almost exclusively broadcast in the fall and spring when the audience is planted firmly on their living room couches.

Second, episodes are delivered with a regular schedule: 22 episodes on the same night every week.

Third, television content is either delivered with commercials or by subscription to pay for the development costs and hopefully earn a profit. The subscription model only works for channels like HBO that deliver a steady stream of both original programming and movies–not the model that games can emulate. The advertising option allows the content to be delivered free to the audience–this is the only viable model for games. The audience-at-large won’t pay for episodic content. Not even a dollar an episode. It’s simply too difficult to get the consumer to overcome the inital cost of buying into a new concept.

Fourth, television episodes have a fixed cost per episode. It will vary for different series and is certainly subject to over-time costs and renegotiations with the series stars, but it is certainly predictable.

How do you deliver episodic game content in a reasonable time frame?

• Build a stable and reliable engine first. The investment in man-hours has to be committed up front. There is no time to develop the engine during development. You certainly wouldn’t rebuild the cameras or lights before shooting each TV episode.
• A very high-level scripting language. Anything that is significantly more complex than a shooting script is too cumbersome to develop content in a reasonable timeframe.
• People only. The time to rig a new skeleton is not a reasonable when developing episodic content. If the protagonist must have a dog, do that up front. And make sure that the rig can be shared with cats.
• Create content for what already exists. If you want to have a UFO appear, plan it out two or three episodes in advance. There is no time to build and test the UFO before “shooting.” You must have a library of animations, props, and sets created well before they are needed.
• Animations and other assets must be shared. It’s okay if most of your characters walk the same at the beginning. If the series catches on, you’ll have plenty of time to develop unique characteristics. Look at the first two seasons of Seinfeld. They are almost unbearable to watch. It wasn’t until the third season that the characters settled into their personalities.
• Hire quality, professional writers. Nothing makes a game worse than playing through porrly constructed plots with stilted dialogue.
• Hire quality, professional voice actors. If you are going to compete for eyeballs with television, you have to hit the same bar of quality. Text on screen simply won’t cut it.

A simple challenge? Absolutely not. But if you can make it stick, you’ll have the pipeline in place to develop multiple series simultaneouly.

Back when I worked at New World Computing, I overheard a conversation between two our lead designers.

“So, how do we want to handle modifiers to ranged attacks?”

“Between 0 and 50 feet, let’s have no modifier, from 51 to 75 a -1, from 76 to 100 -2, and beyond that a – 3.”

“That’s good, but let’s extend it out to 200 feet.”

This illustrates a basic misunderstanding of designing computer games. With this approach why design the game at all. Why not just copy the charts from the Dungeon Master’s Guide verbatim? At least then the game mechanics would have undergone years of play-testing.

These designers failed to grasp the biggest differences between playing a game on the computer and playing a paper-and-dice game.
First and foremost, is the failure to realize that the computer is very good at math, so let’s use it to do that. The “to hit” modifiers are attempting to recreate an exponential curve that has been coarsely quantized to make its application to die rolls more comprehensible. Why not just use an inverse-square falloff?

Secondly, fact that a computer has a physical input device, whether it ba a mouse or a control pad, eliminates the need to have simple random chance to determine success.

So how do we find designers with an ability to “think outside the box?” How do we evaluate the designer during the interview stage of hiring?

I humbly submit that we test them. Tests are routinely applied to programming and art candidates. These test are both quantifiable and subjective. Did the applicant complete the test? What approach was used? How well was the final product executed.

Here is the test I would require of designers:

Create an original multi-player boardgame with a narrative that has no random elements employed during game play, i.e., die rolls, spinners, et c. Randomization may be used during initial game setup, e.g., card shuffling.

By making it a board game with have placed a parameter on the test that elimates classic card games such as poker and more modern game like CCGs.

By eliminating in-game random events, we have disallowed such games as Backgammon, Monopoly, and even Warhammer 40K.

Requiring a narrative be applied, both allows us to evaluate their ability to tell a story through game play and further elinates games like checkers and chess.

This test requires that the designer develops mechanics that test the players’ skill.

This would be a “take home” test, to be sure, but ask yourself how many of your current designers could handle such a test.

The “polish pass” is something that is paid a lot of lip-service in the industry but doesn’t seem to be taken seriously. Sure, Valve delivers polished content, but Valve spent five years delivering Half-Life 2. Not an option that many of us have. Let’s think of polishing as post-production instead. Post, from the Latin postquam, meaning afterwards. That’s right, after production has ended.

What does this really mean? It means “if ain’t broken, don’t fix it.” How do you define a new feature? If the feature requires any new assets, it’s a new feature. Post-production should be a time when the team can fix bugs. Art bugs, code bugs. Just fix things. Spend the time making you AI really shine. Make sure that the art is consistent and appealing. Get that frame-rate up from 30 fps to 60.

Save the new features for the next title. Is the new feature really that important. I mean, seriously. That new whiz-bang feature going to bump your meta-critic score from 72 to 73. Maybe.

Having the time to make what you have really good will do more to satisfy the critics and consumers than any new features.

Most of the games I play tend to be flash games. My favorites are those from Eyezmaze, which are just plain fun, and Façade, which is taking a chance with something different. I guess that makes me a “casual gamer.” I’m certainly not hardcore.

I’m well aware that the games we, as developers, make today are product. The budgets are too big and the investment too large for the publishers to take too many risks with new projects. For some reason, though, I still care about what I do. I want the project I’m working on to be absolutely as good as it can be. And from time to time, I discover that I’m not alone.

This evening I spent 30 minutes standing in an office doorway discussing how we can make the next project work better. How can we free the artists from their creative shackles and still help them work within the constraints of the technology.

Interestingly I spent these thirty-minutes talking with some of the more jaded and sardonic members of the team. I think it’s this care for the game, this desire to create a product of high quality, that makes us this way. It’s not that I don’t like games and it’s not that I don’t like making them. I just want them to better. I want them to be the best games they can be.

In large part, I think this due to lack of or poor quality pre-production. If, instead, the first 4-months were devoted to rigorous pre-production, the team might be able to hit the ground running. If the core technology was in place: shaders, AI code, primary assets—then the rest of the project could be devoted to generating content.

Documentation, too, is key. Every artist and designer on the team should be able to reference detailed descriptions on each piece of the puzzle. Clear instructions on how to construct materials, target polygon count and texture resolution, memory budgets, scripted sequences, the export pipeline, ad nauseam.

Documents are boring to write. People resist “early optimization.” No one is willing to commit to a features lock-down. Mostly there is no pressure. It’s the pressure at the end of a project to hit the milestones of alpha, beta, and release candidate that get the team to focus. It’s the poor framerate and lack of available memory that get the team really focused on managing the code and assets. That cool piece of technology that was supposed to make the game really stand out? It turns out it doesn’t work and you two weeks to fix it. The team gets cut crazy. Features are cut. Content is cut. The project is a shadow of its former ambitions.

So the question is, how do you put pressure up front? Early-alpha? Hard deadlines? How can you be rigid without stiffling creativity?

Maybe we realize that we aren’t that creative in the first place? We reward cleverness, not creativity. In the end we stamp out the same game as last time with a face-lift. Sounds a lot like shovel-ware.

Maybe we base the whole product on one good idea and place all the focus there. Something to differentiate the title from everything else on shelf. Sounds awfully risky. And we all know that today’s publishers are terribly risk-averse.

How do you get the team excited, or at least not indifferent, to the boring parts?

I drew up similar timelines to those seen below on my TD’s whiteboard and didn’t get around to erasing it. The next day he mentioned that everyone who came into his office that afternoon asked about the timeline and not one person was able to disagree with it. This included producers, those responsible for the 12-hour day.

Let’s take a look at the typical 8-hour work day(8 hours + 1 hour lunch):

You’ll see that we can expect 6 hours and 10 minutes of total productivity.
8-hour Day

Now let’s take a look at the typical 12-hour work day:
12-hour Day

You’ll see that by increasing the workday by 3 hours, we’ve actually decreased the productivity to 4 hours and 30 minutes. A drop of 2 hours and 40 minutes.

The only explanation for this drive to increase hours: it’s the only option that production can exercise that makes it appear as if they are impacting the project. Unfortunately, it impacts the project in a negative way.

• Microsoft’s XBOX 360 lead in the marketplace
  Among next-generation consoles Microsoft currently has 99.9% of the marketplace, due entirely to their one-year headstart entry onto the playing-field. Regardless of superior technology, the advance lead into consumers homes counts for quite a bit. The original Xbox was technologically superior to the PlayStation2, and yet, spent its entire life cycle playing catch up to the PlayStation2. Had both consoles been released at the same time, Microsoft would have surely made significant in-roads in terms of marketshare. It would be a mistake to ignore the PlayStation2’s backward compatibility with the PlayStation. As a result the PlayStation2 launched with a large library of titles already in place. With this newest generation, this will not play as large a factor as the Xbox 360 already has a significant library in place and this continues to grow–especially with the Xbox Live Arcade.
• Xbox Live Arcade
  Much like Valve’s Steam, Xbox Live provides a direct link into the consumer’s home and provides a low-cost avenue to provide new titles at significantly reduced price. Unlike the PLAYSTATION3’s target of hardcore gamers, the Xbox 360 has not forsaken the casual gamer–nor has it ignored the retro-game market. The Xbox 360’s consumers adopted en masse classic games like “Street Fighter II” and new retro-style games like “Geometry Wars.”
• Price
  It would be a mistake to ignore price in this equation. Both Microsoft and Sony are offering two flavors of their consoles. The Xbox 360 is available at two price points: $300 and $400. The cheaper of the two lacking only a hard-drive and wireless controllers–each of which can be added on. The PLAYSTATION3 is available at price-points $200 higher: $500 and $600 each. Not entirely a fair comparison, but let’s not forget this is the same price point as the failed 3DO console.
• Blu-Ray
  Sony appears to be betting the farm that consumers will flock to the PLAYSTATION3 as a comparatively inexpensive entry into the Blu-Ray market (stand-alone players currently cost just shy of $1000).
• Developer support
  Let’s face it, developers hate working on the PS3. Much like the PS2, Sony has chosen to make hardware differs significantly from the Wintel world in which most developers are comfortable. For better or worse, Microsoft provides mature libraries and SDKs and excellent developer relations. Sony, on the hand, has notorious bad Dev Rel and offers sketchy, at best, libraries.
• Launch numbers
  Sony missed their promised number of consoles available by a long shot. Europe was delayed and, even after continually revising their numbers down, Sony only delivered 80K out of 100K for Japan and, by most reports, delivered about half of the 400K for North America.
• SIXAXIS controller
  The tilt-sensitive PS3 controller is just a silly copy-cat of the Nintendo Wii’s full motion controller. In addition, since most titles are developed for both the 360 and PS3 simultaneously, the support for the SIXAXIS controller will be a spotty afterthought. Also the lack of rumble support seems to have hit a nerve with the gaming community. I’ve never been fan of controller rumble, but it is puzzling to see an existing feature replaced at the last minute.