Monday, July 12, 2010

Case Study: Prototype Game Development

I've spent a decade doing graphic designs for the game industry. It's generally fun, playful work and with few exceptions the daily decisions are rewarding and creative. When you've worked on dozens of products you are eventually going to get one that just doesn't work out. This case study explores the early creative developmental steps of a game that was never produced.

The inaugural developmental team started small...the art director, (myself), and an R&D lead. The R&D lead is essentially the product owner and is responsible for all gameplay decisions. In other words it is their job to make the job fun to play. My role as art director is to make the game fun and exciting to look at and easy to understand.

The "mission" to create the game was to create a game that utilized some new kind of material, form, or introduced a new conflict-resolution mechanism. All three would be best. The game we were to work on basically was a build-and-battle type game where players collect and build small models that can be used to compete against each other in small strategic skirmishes. This "mission" and decision to create a particular type of game starts with the R&D person noodling around with an idea. They present the idea to the VP of R&D, usually in a semi-casual meeting, and together they decide if the game is worth taking further. This game passed that point and that's where I, the art director, was assigned to the project.

This was a brand new game using untested materials. We knew that we wanted to create a game that had players building small combat vehicle models out of parts they collected then doing battle with their friends. The R&D lead knew that there would be several components to the game: The vehicle, the weapons, and the weapon's unique "die."

We also knew that we wanted to create the game using minutely cut plastic sheets similar to credit-card material. The model sheets would be punched out of the sheet and assembled. This would allow the player to restore the model to the sheet for easy storage, (if desired), and easy to package because the unpunched models were flat. The working title of this game was Punchbots.

We started by creating a creative scope. We knew that we needed some vehicles, some weapons, and some dice to go with the weapons.

At first we thought we would use plain plastic cubes and supply stickers that were unique to each weapon so that one weapon might have certain effects on its die and another would be completely different. This was soon ruled out as we realized how easily it would be to hack the dice to improve the results. Imagine, for example, that you bought white unpainted dice that came with a sheet of black pip stickers. You could certainly put 1 pip on one side and 6 on the other...but you could also put 6 pips on all sides if you bought two packs of the dice.

We then decided to make each weapon's die as cool as the weapon itself. This sounded more fun for the players and more interesting creatively.

The vehicles are the weapons carriers. The vehicles also have characteristics that influence game balance. Some vehicles are light and fast but lack defense. Other vehicles are better armored but less nimble. It's a common balancing theme in vehicle games.

We needed ways of expressing all this stuff. I worked with the company to recruit the design help of a graphic designer on staff that I enjoy working with. Scott and I had worked on lots of projects and I knew he was the right person for this. We got started with some sketches of models that you could make out of flat die-cut plastic pieces.

For the first few weeks we burned through sketchbooks by filling pages with small model sketches like these. Several times a day we would meet and compare drawings and flag the ideas we felt were viable and worth exploring further.

As I drew I realized that the die model should reflect the forms found in the weapons because each die was associated with a particular weapon. While the models would eventually have graphics printed on them, it made sense to have the actual forms be reflective as well.

From that point we began drawing weapons and dice together.

We also quickly realized that the dice didn't need to be square. In fact it opened up lots of game balancing opportunities if some of the weapons had better and/or worse odds of obtaining a particular result.

To complicate matters we associated the number of pieces used to create the weapon model to the number of "hit points" of the weapon. In other words, if a weapon was constructed with three pieces, (like the beam weapon above), it meant that the weapon could receive three points of damaged before if was destroyed. This limitation made the task of designing the weapon forms slightly easier because it constrained our decisions somewhat. It also introduced an interesting strategic decision opportunity to the future player; do you choose a weak weapon with lots of hit points of a powerful weapon that is vulnerable to damage?

You can also see in the sketch above that the beam weapon has a simple tumbler die that can yield four results. (The results on each weapon's unique die would be incorporated in the graphics later.)

While working on concepts for the dice we figured that any kind of model that could produce a random number from a fixed set of possible results would work for our purposes. The weapon above was conceived as a gun that shoots sharpened metal discs. The gun is powerful and only has two hit points; the main weapon "chassis" that pegs into the vehicle and the slot that holds a second piece that resembles a metal disc.

The die for this gun is composed of five separate pieces. When assembled the middle disc can be spun like a wheel-of-fortune. The wheel comes to rest and the number indicated over the pointer, (incorporated into the model shape), is the result of that attempt to use the gun. In practice a player might say, "I'm going to shoot your guy with my Disc of Doominator!" The player then picks up the Disc's spinner and gives it a whirl. The disc comes to rest on a result...say, 2 points of damage...which is then resolved before play continues.

We knew that some dice could be ordinary dice are. Tumblers are thrown and the result is shown when the object has come to rest. We also saw an opportunity to use spinners, like the disc weapon. The other category was a top. It's like a spinner but the entire object is spun rather than just a single moving piece. For the combat shotgun-styled weapon, (above), we gave it a top-like die. The top is spun on the table and when it stops it falls to one side. The raised "pedal" is the result of that attempt. The spindle portion is created using two inter-slotted pieces. The piece with the five pedals slips over the top snugly.

Another spinner we developed was similar to a roulette table. The base is created with two inter-slotted pieces. The spindle is raised flat. One of the base arms is raised to create a sort of pointer. When platter is inserted onto the spindle and allowed to spin freely.

Some of the tumbler-style dice were fun to figure out. For the gun styled after a revolver we developed a tumbler die that could produce 8 results, two of which were highly unlikely (and thus had very favorable results). The tumbler was a "solid" hexagonal column using simple tab-and-notch snug fits.

We knew that the vehicles would also have hit points. Faster vehicles would have less hit points than slower vehicles. Our first line of development was to reflect this as "shields" that were shot off over the course of battle. While the concept had thematic merit, it seriously limited our ability to be creative with the vehicle design as they would all need to have some kind of removable armor piece(s).

Instead of armor plates we looked at smaller pieces that could be removed one at a time like a counter. The first viable results was a spike concept where each spike represented a particular number of hit points. As the vehicle is damaged the spikes are removed appropriately.
We realized that the nature of the material we were intending to work with wouldn't allow these spikes to be small enough. Even if they could be, they would be so small that they would certainly break or be lost almost immediately. The armor and life-token concepts were challenging. We decided that the best direction would be to create a pointer device similar to the spike that could be aligned to a particular printed result. As the vehicle takes damage, the pointer is rotated to the new amount of remaining armor.

We eventually had concepts for 16 different weapons, 16 unique "dice," and 8 different vehicles. We began working closely with our in-house 3D modeling expert. He took our sketches and translated them into flat drawings in a CAD program. This would allow us to make precise adjustments and ensure that the models would work.

Throughout the course of his input we made refinements to the models. This was typically done on print-outs of his models.

Eventually we ended up with models we liked. They were then optimized to fit onto the particular size of plastic card we intended to use. Below you can see the shotgun-inspired weapon on the right side of the card. The weapon "chassis" looks like a pump shotgun with a nub on its base. The nub fits into a hole on the roof of the vehicle. The pronged piece next to it slides into the back of the weapon chassis to create a model that resembles a shotgun with three barrels. Awesome.

The two smaller, pointed pieces are the spindle for the top. The large pentagonal piece fits over the two interlocking spindle pieces.

The left side of the card above has another weapon and its associated die. In this case the weapon is a small machine gun. It has three hit points; the chassis, a small laser site-like piece, and a clip of ammunition. When assembled the weapon looks like a tiny machine gun. Its die is a tumbling cube; the most traditional die in the game. When used in the game, the gun delivers one of six possible results.

The vehicle models were developed in the same way. The vehicles did not need dice but the models needed to reflect their relative characteristics, (e.g., fast and unarmored or slow and armored). Below we see a "Crawler" vehicle. Note that this schematic was created while we were still trying to resolve the spike issue...and that's why you see a number of spike "hit point counter" pieces on the sheet.

Having the pieces all in a 3D program allowed us to fully visualize the model form and proportions. These digital assemblies will become handy in the instructions later.

It was time to start working on the graphics. We called these skins because we realized that we could create different skins for different models to provide more flexibility and decisions in the game without breaking the bank on expensive new die cuts.

I started this process by collecting appropriate visual reference. The internet is awesome for this; you can find a picture of just about anything. The collected visual reference was collected into a series of books. Scott and I would refer to these frequently and they grew over the course of the project. Below is a page from one of the books covering distressed and weathered metals.

The flat models were imported into Illustrator. I felt that because the models were built using flat, white plastic pieces we would never create a deep a vector approach to the skins seemed most suitable. This would also allow us to include lots of little visual details...easter eggs for the attentive. (The rejected consideration was a photographic approach.)

We tested the graphics by adhering large color copies of the skins to light board stock and cutting out the pieces. This particular model is about 5 inches long.

Note that the model above has four slots on the roof. These are the apertures for the weapons. This vehicle provides space for four weapons. (At first different vehicles had different numbers of slots but eventually all the vehicles simply had four slots.) Also note the dial at the rear of the vehicle for measuring vitality. The spikey piece is rotated so that the long rib points to the appropriate number on its life meter.

While this was going on I was also working on other game elements.
Logos were an important early component as it would be used in presentations to upper management, early distributor marketing, and other places. We didn't even have a name yet so I worked on some logo concepts using the working title. The thumbnail sheet above contains logos that try to convey the interlocking panels quality of the game. (We identified early on that we would be proud of this characteristic rather than trying to distance ourselves from it.) There were dozens of these thumbnail sheets created.

Other materials are also being developed at this time. The rulebook and packaging are being developed so that I am provided the maximum time to work out the various considerations. Rulebooks are critically important to a game's success; if people can't learn to play the game, they won't buy it. The rulebook generally starts as a series of thumbnails to establish an overall scale and pace. (Shown below.) When text is received from the editing team, the sketches are translated into actual design documents.

Packaging is being developed simultaneously. (Unfortunately I don't have any sketches or scans that are suitable to use in this case study.) The packaging concepts, like the logo, were intended to reflect the card-like form of the models.

The graphics and models played an important role in communicating the vehicle's gameplay characteristics. The vehicle on the left, (above), is faster but has less armor than the vehicle on the right.

We produced a full set of skins for all the vehicles, weapons and dice. These were sent to a special output agency that could print then cut the pieces out of the appropriate stock. At this point we were about 6 months into the project. When the final pieces returned we assembled some extras to ensure they fit correctly. Below we see one of these "actual size" models against the hand-made mockup.

Below are several of the vehicle models together.

Each model fits neatly in the palm of your hand. There were a total of 8 different vehicles. We created two skins for each vehicle...thematic variations. For example, one skin might be a "friendly" character and the other version of that vehicle might be a "tough" character. This allows the player to choose their vehicle partially, (or fully), on an emotional basis if they desire. This is an important impetus for younger players unprepared to make decisions on mechanical and game-balance calculations.

When we finally had our mockup product prepared and the rulebooks written and laid out, we collected groups of kids to try the game. The purpose of the focus group was to witness first-hand the intended market's reaction to the product.

(Some people get "focus group" confused with "play-test." A play-test is done with full disclosure with a desire to identify weaknesses in the game. During a play-test the developer will often sit with the players and coach them through portions of the game that don't require attention, (or that require special attention). By comparison, a focus group uses participants who have no idea what to expect. They are not coached in any way. The developmental and marketing teams will usually sit behind a mirrored wall and carefully observe everything that goes on or record precisely what is said. During a focus group, overall reactions to the product are more important than gaining insight into a particular aspect of it. (Sometimes wealthier companies will have focus groups for very specific and isolated changes in their product such as a new scent, packaging color, shape, or logo.)

For this game the focus group attendees unanimously hated the game. They didn't want to build the models. (Model-building is a lost activity.) They felt like the line between "build" and "assemble" was too blurry. Some of the models are complicated and although we took pains to create build instructions that didn't require any reading, a majority of the participants simply tossed the instructions aside immediately...then became confused...then frustrated.

After the focus groups we decided not to produce the final game and were assigned to new projects.

That's how goes.