Wild Shadow Studios

Ad Hoc Grouping

Early MMO designers wanted people to form groups. After all, why play an online game, if not to interact with others? And what better way to interact than in a group, just as people had been doing for years in MUDs and D&D?

The problem was that those early MMO designers tried to encourage grouping by making it painful to solo. No single character had enough hit points, healing, or damage-dealing power to make much progress alone; only a group could adventure effectively.

But of course grouping is a pain in the neck. To join a group, you first have to find a nearby group of the appropriate level, with an empty slot, which needs help from someone of your character class. Once you've found a group, you have to risk rejection from a stranger by asking the group leader to let you join. And once you've been let into a group, you have to deal with immature players who call you a noob while making you wait until they finish their pizza before they ninjaloot the boss and log off.

In 2008, Warhammer Online introduced the public quest: a recurring combat challenge in a given location. Anyone that wanders by may join the public quest without asking permission, and everyone shares the rewards automatically. This mechanism is great for painlessly pulling people into co-op play, but it suffers from silly repetitiveness (the mummy will respawn in 4:59, :58, :57...) and the ghost-town effect of not being able to gather enough people of the right level.

The main grouping dynamic of Steel Chaos is more or less "holy crap we've got to attack/defend location x RIGHT NOW!" I call this "ad-hoc grouping": players working in close proximity on a common cause, but without a formal grouping mechanism. Steel Chaos is massively co-op, so the players are already working on a common cause. The game has neither XP nor loot, so the issues of mobcamping, killstealing and sharing rewards with freeloading strangers just aren't there. And since anyone can play any role and everyone is always the same level, there should be fewer problems gathering people to help.

I want Steel Chaos to break down the barriers separating MMO players from their friends. My hope is that ad-hoc grouping will painlessly provide the social benefits of grouping without the problems that normally attend formal groups or public quests. 

Easy Class Switching

In most MMOs, the first thing you do is create a character. People love doing this of course, picking name, gender, race, body shape, costume, hairstyle, facial features, etc. You also pick a class: a set of capabilities which unlocks over time as your character progresses through the game. Of all the choices a player makes when creating a character, class has by far the biggest effect on the gameplay experience.

The problem with choosing a class at character creation time is that it locks the player into a certain gameplay style for the remainder of the game. A new player may not enjoy the gameplay of a certain class, but she doesn't know that yet. What if she chose a warrior class expecting to slice up herds of enemies like Conan, but it turns out that warriors are just meat shields that can't really fight that well?

Even worse, being locked into a class can prevent you from playing with others: if you are a tank and the group already has a tank, you can't play with them. If you are a wizard but the group needs to fill its last slot with a healer, you can't play with them.

Another problem with permanently-chosen classes is that the MMO designer must constantly adjust and fine-tune the balance between classes, to make sure that all the classes are of equal power. Otherwise, a player who inadvertently chose an underpowered class would have a poor experience. But the buffing and nerfing of the balancing process also annoys players, who feel tossed about by the whims of game designers.

To address these issues, Steel Chaos allows players to change "class" anytime. In Steel Chaos, the players are hovertank pilots, and their in-game capabilities are determined by the kind of tank they're driving at the moment. To switch class, a player just chooses another tank; any player can drive any type of tank.

My intent is that no one will get locked out of a group because of class. If the group needs another repair tank or additional heavy-assault capability or more crowd control, anyone can perform the needed role by jumping into an appropriate vehicle.

Steel Chaos

GameTwo has a new codename: Steel Chaos. I just ran a multiplayer playtest, and I'm looking for more playtesters. If you'd like to help, drop me a note at rob@wildshadow.com.

Procedural Content Generation

Traditionally, MMO game worlds have been lovingly hand-crafted by an army of level designers, texture artists, character modellers, animators and quest writers. These people costs tens of millions of dollars, and the resulting games occupy many gigabytes and require hours to download, install and patch. I'm looking to avoid most of this for GameTwo.

The main technique I plan to use is procedural content generation: having a computer program create the world. This method has been used many times in the past, in classics like SimCity, Civilization, XCOM, and Age of Empires. These games used algorithms to create game maps, and I plan to do the same in GameTwo.

But I'm also working on taking procedural content generation a step further, by creating new enemy types on the fly. As the game progresses, the enemy forces design new vehicle types in order to be more effective against whatever weapons and armor the players are using the most. Using a parameterized drawing routine, GameTwo creates new graphics on the fly for each design.

Each enemy tank design has different weapons, different hit point totals, and resistances to various damage types. The goal is to have the enemy adapt to what the players are doing, challenging the players to devise new equipment and tactics to beat them.

Cooperative Leveling

My biggest goal for GameTwo is to create a game where the players can have a real effect on the game world. The idea is to get away from the static worlds of traditional MMOs and move to a model where the players shape the game world every time they log in.

GameTwo's players are hovertank drivers in a cooperative PvE war. The world has been taken over by enemies, and the players' goal is to win the war and retake the world. This is a big task, which may require months of effort. The players start out weak, in a constrained area, beset by enemies on all sides. They must push the enemies back, struggling to secure the key resources that will let them continue the war.

A typical play session involves joining a group of players to fight the enemy, either offensively by trying to take over territory, or defensively by protecting land from attack. The game is pure sandbox; there are no quest hubs or missions. The players decide where to attack or defend, and they live or die by those decisions.

A big difference between GameTwo and traditional MMOs has to do with levels. In a normal MMO, each player levels on her own, tackling harder and harder content in level-appropriate zones. However, this system won't work well in GameTwo's open-world sandbox, because there are no zones. Everything is one big world; everyone fights the same enemies.

The approach I'm trying is to have all the players level up simultaneously and cooperatively by researching better equipment, such as weapons, armor, and engines. Once something is researched, it is available for anyone to use. The research is paid for with resources extracted from the world; the idea is that the players will work together to secure resources, which are then spent on research projects to improve the players' capabilities.

Research is democratic; each player gets a voice in what technologies will be worked on.  This dialog shows some research projects in progress:

Each player can direct her share of the resources to whatever research project(s) she deems worthy. Projects with a lot of popular support get more resources, and hence they complete more quickly. In this way, the players chart their course together, working on the technologies that they think will help their collective war effort the most.

Parallax Scroll

Last week I was experimenting with parallax scrolling in GameTwo. Parallax is the effect a moving observer notices when viewing objects at different distances. In this case, the roofs of the red and blue buildings are closer to the viewer (the overhead camera) than the ground is.

After I move my tank to the left a bit, GameTwo draws the roofs in slightly different positions relative to the building footprints. This pseudo-3D effect is subtle, but it makes the buildings seem much more real.

I'm thinking of using this parallax effect for lots of things: rocks, buildings, trees, mesas, canyons, crevasses. Can you think of other uses?

Vehicle Editor

Yesterday I built a rudimentary in-game vehicle editor for GameTwo. It's a dialog that pops up in the game client, with sliders for changing the various parameters to the tank-drawing code.

When finished, this tool will help me generate new graphical assets quickly. A more advanced version could be used by players to customize their tanks.

Dreadnought

After a playtest session with a friend today, I decided to add a much tougher enemy tank to GameTwo. The new Dreadnought class tank has three times the hitpoints of the previous heavyweight, the Destroyer. It's got two powerful laser beams, one in its turret and one mounted on the bow. It also has a missile launcher which fires a superfast orange homing missile every two seconds, with a 180-degree forward-facing firing arc. Standing in front of this thing is not a good idea.

The graphics for the Dreadnought and its missile were easy to add, using the procedural art system I built last week. But a tank this size could use a few more details. Maybe some antennas?

More Procedural Art

I did some more work on GameTwo's procedural graphics, replacing all the hardcoded enemy tanks with new versions created at runtime by the server. As with last week's missiles, the new tanks are all drawn by the same piece of client code, based on parameters provided by the server. The parameters include:

hullcolor, treadcolor, turretcolor, width, length, nosewidth, noselength, tailwidth, taillength, treadwidth, treadlength, treadlateraloffset, treadlongitudinaloffset, muzzlewidth, breechwidth, barrellength, turretwidth, turretlength, turretlongitudinaloffset, turretnoselength, turrettaillength

These parameters yield a decent variety of enemy tank types, but I look forward to expanding the available parameters and adding new vehicle types as well. Maybe planes?

This screenshot shows my tank getting completely worked by a screenful of the new baddies. The red lines are the enemies' lasers, caught in mid-animation. The mostly vertical laser was fired from the large tank at top center. The other two beams haven't hit me yet, but they are about to finish me off.

The smallest enemy tanks are the Fighter type, which have a short-range turreted laser and the ability to fire bright red homing missiles. The medium-sized grey tanks are the Cruisers, which sport front-mounted medium-range lasers and fire yellow grenades from their turrets. The large Destroyer tank has a powerful long-range turreted laser, in addition to hull-mounted rockets and grenades.

The dark red enemy tanks are a new type (Gunship) which launch large numbers of twin-nosed homing missles. The new procedural art made it pretty quick and easy to get both the Gunship and its missiles into the game. I'm hoping that the art system, when fully deployed, will let us rapidly insert many new types of enemies, allowing the live team to evolve the game over time in response to player actions.

Procedural Ammunition

The art for GameTwo is 100% vector based -- there are no predrawn bitmaps. I like things this way because procedural vector art is cheap to produce, extremely space-efficient, and easy to modify. However, the vector art in GameTwo has always been hardcoded into the client at compile time. This unfortunately makes adding new game elements (such as a new weapon) more painful because both the client and the server must be modified: the server for game logic, and the client for the art.

I recently spent some time giving the server the ability to create art assets at run time, without rebuilding the client. The first asset I converted to the new system was the homing missile. This missile had been just a simple blue pentagon, which was really confusing because the enemies used them too. It wasn't always clear which blue pentagons were yours and which were out to kill you.

I wrote some code in the client which draws a missile based on some parameters: color, width, length, nose cone shape, fin size. To generate a new missile, the server sends the parameters to the client. Now when I'm coding up a new type of missile in the server, it's easy to add the missile-drawing parameters at the same time, without having to mess with the client at all.

This screenshot shows three types of missile objects, drawn by the same code in the client, based on parameters supplied by the server.The fat blue missiles are the homing missiles I've just fired into this enemy town. The red missiles are the homing missiles fired at me by the small enemy fighter tanks. The skinny orange finless rockets are unguided missiles fired at me by the flower-shaped alien buildings.

I want to deploy this system for all of GameTwo's art: tanks, ammunition, buildings, even ground clutter like rocks and other doodads. I think it will give me the flexibility to quickly deploy a wide variety of interesting game elements.