Game Design

Captain Forever Wins IGF China!

The first of many plaudits is draped on Captain Forever. All is right in the world.

We proselytize about this game often; a few words about its design.

In Captain Forever, the player constructs a ship.  That ship is then tested against algorithmically generated ships that drift slowly through the game’s infinite playfield.  Scavenging the parts of destroyed ships, your ship grows and evolves. There are ten levels of girder, thruster, and guns – on a scale from green to white.  The game always sends some ships of a much higher level  into the fray, providing that oh-crap-time-to-run kind of feeling. Green is prey, white is terror.  As in Spore, skillful play enables you to later be that massive ship, fearlessly preying on smaller and weaker ships. The feeling of mastery over previously mighty enemies rarely so delightful.

As a soundbite, Captain Forever is what I wanted Spore to be. Not the intricate simulation, but what most excited me about Spore: meaningful construction. The placement of parts on your ship has a deep, satisfying impact on how effective that ship will be in the ecosystem of the game.  It also has an immediate, visceral impact on the feel of control. A well-constructed ship both feels good to control and is effective in the game. It’s a perfect marriage of construction and utility, a perfect balance between skill and construction.  There are so many strategies and so many ways to build an effective ship.  Depending on your construction, you may adopt a new style of play. Or you may see a particular ship you want to attack, and quickly rebuild to a more optimized battle configuration.  It’s so fluid; building happens in the middle of battle.  The marriage of construction and play is seamless and simultaneous.  The game’s ecosystem is perfect in conveying meaning to your construction.

The game is an epic moment generator. As you fly around, there is running internal monologue: can I defeat that ship? Will it be worth the damage I’ll sustain? How does that ship stack up against my current construction?  Can I destroy that ship without damaging the parts I want? What will my ship be if I get those pieces?  Sometimes an attack on a massive ship many levels higher than you will result in the destruction of most of your ship. But sometimes you’re able to worm your way into the innards of that ship, a parasite, and chip away at its core until it succumbs.  A fight might take five minutes to complete. And then a mad scramble to rebuild before the next massive enemy frigate rolls through.  First girders, then steering jets, then a gun of some kind. You must always be ready to zip away, and that means having a balanced controllable ship at all times.

It’s also possible to ‘kite’ two large ships together. Higher level ships ignore the puny unless provoked. Provoking two ships and running them into each other to wear them down a bit is obscenely satisfying. It’s just so good.

Nuts and Bolts

There are hundreds of small, brilliant decisions that make all this work. Many of them are tuning: Box2d is used masterfully here, and each level of girder, and gun, and thruster is immaculately tuned to its own level and to the other levels. A purple thruster on a ship made of orange pieces means crazy speed. Orange thrusters on a ship made of purple girders leaves your ship pathetic, unmaneuverable.  Guns are relatively light, but they need girder space.

Construction is simple and intuitive. As you click and drag a piece around, it tries to orient to the nearest attachable surface. It does this just slowly enough to allow you to fiddle pieces to just the position you want. I still occasionally pull off a piece unintentionally, but I always feel as though it’s my fault; I was hurrying too much and slipped up. This is perhaps the most difficult interface problem in this design, and it’s been solved masterfully. You don’t think about it, you just put the pieces where you want them. No small design feat.

Another brilliant design cluster is in the notion of ship “hearts”, and the tree branch model of damage and destruction Farbs employs.  To win another ship’s pieces, you must destroy its heart. But its heart is often its center, the main note to which all other pieces are attached. Ipso facto, it’s difficult to shoot, especially without destroying other parts. But the whole reason you attack a ship is to get its parts. This tension, between desire and danger, is exquisite.  Likewise, when a piece is destroyed, all pieces attached to that piece are also destroyed. This forces even greater caution and though in one’s angle of approach on a potentially lucritive enemy target.  You can’t try to blow off just a wing and then salvage the bits that break loose: you have to go whole hog, all or nothing. You have to destroy the ship to save the wing, with all its juicy gun and thruster goodness. This is just plain old fashioned awesome game design. Pure, elegant, excellent.

This game tickles my brain in all the right ways. It is, as so many other great games have been, a recombination of familiar elements.  In the relationships between these elements, Farbs has found something new, beautiful, and elegant. A well deserved win for Farbs and his amazing game. Cheers.

GO PLAY CAPTAIN FOREVER RIGHT NOW

To people, time is linear.

Memory allows us to revisit and relive the past, to learn lessons from it, and to develop different attitudes and ways of behaving in the world based on those lessons. But memory is an imperfect recording, heavily influenced by emotion. Two people can remember the same party as being wonderful (it was the night I met my future wife) or awful (it was the night I walked in on my wife with another man.)  Objectively, the details are the same. The party happened at a certain place, at a certain time, and included specific people. But there is no way to separate emotion from memory. All memory is inherently subjective. The arrow of time cannot be reversed. But what if it could?

It is tantalizing to imagine what the world would be like if time were malleable. This is the essence of time manipulation in video games. One of the appealing things about video games is that they allow expression of our desire to exercise control over time.  A game allows us to escape the way we perceive time by rewinding and replaying, collaborating with past sevles, or simply hitting the reset button.

I’ve been thinking a lot about time manipulation in games lately because our next game, Time Donkey, features rewind/replay mechanics. Here are some thoughts on the design of other time manipulation games I’ve been playing. For reference, of course :).

Cursor * 10

Cursor Times Ten distills the essence of multiplicity. The interface is a cursor. The only novel idea in play is the recording and playback of in-game actions. Everything you do is recorded. By careful planning you can collaborate with all your previous selves to achieve objectives more efficiently. What’s most interesting is comparing your memory of what you did before to a perfect, objective one.  I feel as though I moved my cursor over as quickly as I could, clicked the stairs, and then went on to click the button as many times as I could as quickly as I could.  But the replay shows that I over-corrected my initial mouse movement, clicked the wrong thing twice, and only managed to click the button ten times before being forcibly restarted.

What Cursor*10 really nails is the sense of upward progression. There is a secondary goal of scoring, but what feels compelling is clever application of multiple replays to optimize tasks.  Once you’ve revealed the box hiding the staircase, you can quickly and easily find it next time, allowing the replay after that to reach further and further up the floors. This sense of building intent across multiple plays, and interacting with your own previous actions feels fresh and interesting.

DefeatMe

DefeatMe is decribed in comments as “Cursor*10’s evil twin.”  This is an accurate statement.  In Cursor*10, you replay the same small slice of time over and over again, layering recorded replays one on top of the other. In DefeatMe, your replays are mirrored opposite, and become an increasingly wily and difficult opponent.  Every shot you took at an enemy comes back at you, and every move you made to avoid a shot becomes a clever dodge in the next round.  Each dodge is predictable if you plan carefully, and you can anticipate where it will end up and send a shot there before the ship arrives. This makes for a satisfying, cerebral experience (though this is currently hampered slightly by having to reload the page to start a fresh game.)

One weird decision in this prototype is bullet spray randomization.  Your ship starts by firing one bullet at a time, directly forward.  As you progress through each stage, however, the pattern can change. Sometimes you spray three bullets. Sometimes you have no direct forward shot. Presumably, this is to solve the problem of fatigue. If your ship only ever fires one bullet and directly forward, it’s too easy to predict where each shot is going to go and the play gets stale. I’m not sure changing the bullet spray is the right answer. The problem is that it feels arbitrary. It might feel more satisfying and strategic if there were collectable powerups in the level that changed the bullet spray pattern or upgraded it.  Certainly, as the levels progress and there are more and more replay ships to destroy, I’d choose to then upgrade my weapons.

Regardless, this is a lovely little experiment by crafty indie shmupomancer Kenta Cho, one which I’d love to see him expand on. I wonder what it would be like with varied level terrain, powerups, or with the addition of patterned enemies independent of or spawned by the player.

Chronotron

We have a joke around the Flashbang Office:  That’s a Problem for Future Me.

As in: ”Hey this code is unoptimized and barely works. Ahhh….Future Me problem! Present me will eat cheesecake instead.”

or

“Man, I probably shouldn’t have eaten all the cheesecake…ah well, I’m sure Future Me will figure something out.”

Chronotron is neat, but I disagree with one of its self-implied design constraints.  As in Cursor*10 and DefeatMe, your actions are recorded as you play.  To start a new replay, the current robot must be returned to the door from whence it game. I don’t like having to return each replay robot to the door.  It is a fundamental constraint to the levels as they’ve been constructed in Chronotron, but I can’t help but think the game would be more enjoyable if it flowed forward.  I like thinking about how to use multiple replays to activate switches, weigh down scales, or launch off of seesaws. Just let me do that!

Having to then solve all the annoying paradoxes that occur when one robot prevents another from getting back to the door feels tedious. If there were a second exit door, or various portals throughout the level I could enter to start a new replay, and the puzzles would remain in tact sans tedium.  Fixing paradoxes just feels like sweeping up after the clever bits. However, the ability to interact with oneself across time – and the many clever physics-y applications they found – makes Chronotron worth playing and enjoying.

Braid

Braid is pure experimental game design.  Johnathan set out to explore time manipulation because he was unsatisfied with the limited, tepid, unsatisfying application it had found in games like Blinx and Prince of Persia: The Sands of Time.  He asked the question ‘what would happen if the player could rewind time as much they wanted, as far back as they wanted, any time they wanted?’  Layering limitless rewind on a familiar feeling 2d platformer yielded unique and interesting gameplay.  He then explored the mechanic more thoroughly through a world structure that modified limitless rewind with additional constructs like a ring that creates a bubble of slow time and an arbitrary rule linking the flow of time to left or right movement. The results were revolutionary.

This is the finest game about the nature of time as it is perceived by humans, and one of the finest pieces of work in any medium on that subject. There are ideas in Braid that are best expressed interactively. The concepts can be explained linearly, but are better expressed through interaction.

Coming Soon: “The Misadventures of PB Winterbottom”

A quick parting note; add this game to your List of Future Me Things . It’s being created by our friends The Odd Gentlemen, and has the potential to outdo Braid in terms of thoroughness.

Winterbottom will explore replays as help and hinderence, play with the arrow of time, distort space and time in even more ways than Braid. Seek it out!