Geisterfahrer: Rubber Meets the Road

After writing about it several times, I finally had a chance to actually play Sleep is Death this afternoon. I was the Controller (having never been a Player), with another local game designer being the Player. Unsurprisingly, it was nerve-wracking feeling like I had to juggle so many things at once. It wasn't a bust though; we definitely created a cogent story and the experience itself was very interesting. The story actually ended up kind of clever and went places I didn't expect at all. That's successful in my book.

(As an aside, the above photo is from an abandoned prison farm near Atlanta, GA. Yikes.)

I'm going to withhold posting a description of the story itself for now, as I'm very curious about using the same initial setup with other Players and see what things change. I did want to write some thoughts on how the experience went functionally. This could bleed into some more meta/structural ideas as well. Without further ado:

0) The default turn time of 30 seconds isn't nearly enough for new players.

Go to your Sleep is Death install directory, look for a folder called settings. Inside there will be a file called timeLimit.ini, which has only a single value in it that is the length of a turn in seconds. Open it up in Notepad or another text editor and change it. We doubled it to 60 and that seemed far more manageable. This seemed more important for a novice Controller like me. More familiarity with the tools and I probably could have handled 30 seconds, even with a novice player.

1) The available assets will drive where the story goes.

Even with turn time doubled to one minute, this still isn't nearly enough time to create significant new assets on-the-fly. I'm sure the fact that my art skills are rubbish doesn't help, but I couldn't improvise anything but the simplest objects. New characters? New rooms? Forget about it. The story went to a courtroom and a graveyard because, well, there were assets for those rooms available. I didn't plan going there beforehand, but when the locale had to change, the possibilities were pretty limited.

I had created two starting rooms and figured I'd just let the story flow from there. But I didn't realize there weren't many places for it to flow to. The number of locales available in SiD by default are quite limited. I was expecting a handful of "stock" rooms, like an apartment or a store. Not quite the case. The scenes available seem like they were used in other stories and come prepopulated with characters and objects. Should you want only the space, you have to clear out the rooms in dwindling seconds.

My recommendation is that if you want a more freeform story, find or modify a number of generic locations that can be quickly repurposed.

2) Searching for assets could be much better.

The only search functionality for assets in SiD is searching over the titles of objects/rooms. These titles are not unique (multiple objects can have the same name) and have a 10-character limit. Searching for "person," "guy" or "lady" results in nothing. Characters are titled by fictional names.

Tag-based taxonomies are a bit fad-ish, but having some kind of lightweight attribute system would make finding specific types of objects much easier. Even being able to filter by people, animals, interior objects and exterior objects would reduce the time-pressure felt by Controllers needing to find an unexpected object.

3) Adding assets from the community is difficult.

Sleep is Death facilitates adding assets created by the community members. Adding the assets is quite easy, but due to the above challenge with searching, actually utilizing them is less simple. The main community site, SiDTube, has a resource page whose searching is basically the above. It can't be filtered by music/objects/rooms. The scale of some objects varies wildly from the stock SiD assets. There is a rating system, but it's not that useful when assets like this have five-star ratings.

The framework of Sleep is Death means just about any style of sprite-based art can be easily used. The stock resources for SiD use a very basic 8-bit style. Unfortunately, many of the community resources do not. Shannon Galvin created a fantastic world-building tutorial that looks gorgeous, but his resources are unusable unless you're willing to create more assets of the same fidelity (which is well beyond my talents and spare time).

I very much appreciate that Rohrer has structured SiD to easily incorporate community-generated assets. But it seems that, like the in-game asset manager, better tools for searching and organizing assets would make adding community assets much easier. If you're expecting to use community content to expand your story, expect to spend quite a lot of time searching for assets that match your aesthetic. And even then, they will be a baseline for modifications rather than drag-and-drop additions.

4) Clarify expectations with the Player beforehand.

I didn't confer with the Player about our story at all before we began. As such, I think we had different ideas of how the game should progress. I had two initial scenes vaguely planned, but had not (intentionally) structured much beyond that. I did not setup any kind of goal or motivation beyond a vague time limit within the narrative itself.

Coming from a tabletop background, I was very leery of "railroading" the Player. I wanted to provide as much freedom as I could feasibly support. However, the Player was initially expecting more direction from me (we had a little debrief afterward). He kept his statements a bit vague, concerned that he'd say something that would upset or derail the story I had planned. Little did he know there was no planned story to derail! I was hoping he'd commit to concrete details, allow me to play off those in ways he wouldn't necessarily expect. About halfway through he took control of the story, which is what I had expected all along.

In theatrical improv, players can see when other players are floundering and try to bail them out. I'd say that interplay to always keeping the scene moving is actually a vital part of theatrical improv. However, in Sleep is Death, there's no way to ascertain if the Player is struggling to keep the scene moving. Clarifying expectations with the Player and either a) letting them know they're wholly in control of the story or b) provide more specific goals/motivations should help ameliorate possible floundering.

A lot of this may sound critical, but I very much enjoyed this first foray with Sleep is Death. I just didn't realize quite how much preparation was required in terms of assets. Going forward, I'll make a point to converse with Players before the game and make sure I have a good suite of scenes (ideally, generic ones) available.

Again, I'd encourage you to check out Sleep is Death for yourself, especially now that the game costs only as much as you want to pay (and for two copies of the game). And for anyone who has spent time on the Controller side, I'd be quite curious to hear how it went for you.

In Medias Ludus

Continuing with the discussion of games and improv theatre, Michael wrote a great post this week about how the methodologies of improv can be applied to Sleep is Death (and life in general). Given the inherently extemporaneous nature of that game, improv techniques are quite directly applicable. But, as I alluded to last time, while I was reading Truth in Comedy, I was struck by how relevant their observations were to digital games in general.

Michael writes, "A Player needs 3 things: 1) A character who wants something. 2) A clearly defined situation, ideally with an inherent obstacle or conflict. 3) Strategies for getting what you want."

When digital games go off the rails, typically one of these three things is absent or vague. The player's goal is unclear ("I don't know where to go now"), the obstacles read poorly ("Is this room a dead end or did I miss a door somewhere?") or it's not clear how to start achieving the goals ("I can't see any way to open that door"). When I'm banging on about readability, this is what I'm talking about.

Note, this does not mean everything should be painfully obvious. The problem is when the player is at a dead end and feels like they have no means of discovering what to do or how to do it. Having to explore and puzzle things out is good, having absolutely no idea what to do and being reduced to trying things completely at random is not (e.g what happens when you're stuck in basically every old-school adventure game).

The other interesting lesson from improv I wanted to call out is "Start in the middle." Rather than starting at the beginning of some action or narrative, which tends toward unexciting exposition, the scene starts in the middle of some event. This technique is called "in media res" or literally, "into the middle of affairs." It's more common in film and television (e.g The Usual Suspects or Reservoir Dogs) than in games.

Despite my increasing intolerance for their staid mechanics and overwrought narratives, a number of JRPGs including several Final Fantasy incarnations have used this technique to good effect. So did Uncharted 2, both Max Payne games and the recent Tales of Monkey Island, for example. And of course, Planescape: Torment's "You wake up on a slab in The Mortuary" may be the best in medias res opening of all.

Opening in this fashion is still quite uncommon though. I wonder if part of the reason why more games don't open in medias res is that the character's mechanical progress moves with the game's narrative progression. By opening in the middle of some action, it may seem harder to have the character mechanically progress from there (and if flashbacks occur, one might feel like abilities ought to be removed). Many games also open by introducing mechanics in a tutorial-ish fashion, which may be a bit harder to fit into a scene or story already in progress.

A few games also begin in the middle of the game mechanically, but at the beginning of the action narratively. Most notable would be the Metroid Prime series, where that player begins with powerful equipment only to lose it after a short opening scene (I think Super Metroid opened this way as well, but I can't remember for sure). These abilities are slowly recovered until the player eventually surpasses that initial state.

We might call this structure "in medias ludus" (Update: make that "in medium ludum," thanks Roger!) if we wanted to be delightfully pretentious, and to be honest, I'm not sure how successful I feel it is. For me, it often gives the feeling that for at least the first half of the game, you're behind the curve and simply playing catch-up. The Metroid series is so well engineered, and this formula is fundamental to its gameplay, it works. But other games that have attempted this, most recently I remember Prototype doing so, don't work as well.

I generally like in medias res openings, partially because it dovetails well with another improv lesson: "Take the active choice toward forward action." Especially in improv, given the choice between discussing some possibilities and simply doing one of them, always opt for the latter. While Shakespeare can make Hamlet-esque indecision fascinating, neither improv nor games are particularly suited for this. If a game explicitly offers a choice, it ought to be between one type of action or another, not between action and non-action. And by "action," I simply mean something meaningful for the player to do, not "action" like shooting guys in the face.

A great many games have sleepy, unengaging beginnings, rife will text-laden tutorial billboards and mundane tasks. This is something no game can afford, especially for those of us in the digital download space, where demos and conversation rates are of vital importance. But even more fundamentally, opening in media res can help focus a game. Starting at the beginning makes it very easy to justify a bunch of dry exposition about characters and a world the player doesn't care for (yet, hopefully). Starting with action and conflict, something that will be engaging for nearly everyone, we can draw players in. Don't throw a bunch of narrative at the player right out of the gate; wait until questions have been posed before supplying answers.

At the end of the day, our currency is engagement. Structuring a game to start in media res is only one approach. The first five seconds to five minutes of a game are vital, however. All games would be well served by a careful consideration of this portion of the experience.

All that from a short book about improv, eh? Again, I highly recommend it.

Truth, Improv and Games

I read Truth in Comedy this week. If you look past the 90s "wacky font" adorned cover, inside lies a surprising and excellent book (and one not unrelated to games). It's only incidentally about comedy; it's about improv theatre, specifically an improv structure called the "Harold." The Harold was developed by Del Close, one of the premier influences in improv and if you look at his list of students, you'll see he trained an absurd number of now-famous folk.

Coincidentally, Sleep is Death was also released this week. While not exactly improv, Sleep is Death certainly shares some improvisational qualities. I haven't had a chance to dig into it yet, but it seems like it possesses the same semi-structure of tabletop RPGs. I've seen some discussion about Sleep is Death devolve into scoffing "Why not just play D&D?" style comments. But I think the structure of Sleep is Death is different enough to be quite interesting.

The representational layer Sleep is Death puts between the Controller and player both constrains and frees the storytelling. All interaction is mediated by the simulation; you cannot see the Controller nor can they see you. It sounds obvious, but that's a big distinction from live roleplaying that mimics real social interaction. And the dynamics of a single player and Controller definitely differ from that of a group of players.

The best response to any scoffing about tabletop RPGs vs. digital ones was presented by Kieron Gillen on RPS. Go read it.

Unfortunately, much like tabletop RPGs, I can only imagine the quality of the experience will vary wildly depending on who the player and GM/Controller are. And, again like tabletop RPGs, there will be a raft of cliched stories, some extreme awkwardness and a few shining gems. I hope finding those gems won't prove to be too difficult. At least with Sleep is Death I won't end up in the basement of an over 30 man that still lives with his parents. A basement whose walls were lined with Star Trek figures, still in the package (seriously, that's no exaggeration).

Anyway, that's enough carrying on about a game I've barely had time to investigate. Back to Truth in Comedy. Given that improv is all about a group of people collaborating to create something from nothing, it shouldn't be too surprising I found parallels with games. There were similarities both with games themselves and team dynamics of creating games. Given what I wrote last week, I'll make a few observations about the latter and the former in a separate post.

One thing stressed in Truth in Comedy is the absolutel necessity of the group working as a single entity. Agreement with fellow players during a scene is the cardinal rule. It's "Yes, and ..." that all improv is built upon. It's trusting your fellow players will provide the information you need and that they will support whatever decisions you make. It requires sublimating one's ego for the benefit of the group. As the authors succinctly put it, "The best way for an improviser to look good is by making his fellow players look good."

The same clearly applies in any team-based creative endeavour. Granted, improv has a bit more freedom in that its creations are inherently transitory and even the wildest of ideas can be chased. But when it comes to game teams, that same trust is vital. Developers have to be confident that their teammates will not judge them based on how successful an idea might be. A judgmental climate means people will be more conservative with their ideas and with their work. Rarely will this benefit a game's development.

The audience of players isn't going to care which name on the credits was responsible for some particular success or failure. The only evaluation they will provide is of the work as a whole. And thus, everyone must focus on the work of the group as a whole, not any particular individual.

Now obviously, I'm not going to fire up Maya or Photoshop and tell an artist what to do. But rather, everyone should feel like their decisions will be supported. The team will provide honest and non-judgmental feedback, and more importantly, will help in making their ideas real. When everyone operates with the same singular purpose, to make the best game possible, disagreements are possible without a feeling of someone having to "lose." One line I particularly liked from Truth in Comedy was, "Treat others as if they are poets, geniuses and artists, and they will be."

Reading Truth in Comedy, I couldn't help but feel that the attitude it presents is similar to the one I see manifest at events like the Global Game Jam. At game james, four or five have to create something from nothing with little preparation. It's clear that the groups that succeed are the ones that trust each other and work with a single purpose. The ones that don't succeed often bicker and measure egos about the "right" way to do something.

I'm definitely going to be checking out a few Harold-inspired improv shows in the weeks to come, but more importantly, I'll be thinking of ways I can better support my teammates and keep our ultimate purpose in mind. Truth in Comedy really was an interesting read and I highly recommend it to basically anyone.