Three Hours with Velvet Sundown


Last month when I ran my charity auction, the three-hour winner was the team behind Velvet Sundown, an interactive drama that I’d heard about but hadn’t had a chance to check out. Their request was that I spend some time familiarizing myself with the game, then draft an initial pitch for a new scenario.

Velvet Sundown allows players to connect and play through scenarios together: all the scenarios are set aboard a yacht and involve a set of prefab characters, but they have widely differing aims. The episodes so far include a murder mystery and a sort of industrial espionage intrigue, among others. Players are given instructions about their character goals, together with props and sometimes some conversational actions, but they’re also allowed to type freeform input to communicate with other characters. The idea is that you can role-play the dramatic situation however you want, while using an in-game mechanic to make any game-ending-critical decisions. (For the murder mystery, for instance, each character can write the name of another character in the logbook to indicate the guilty party — a bit like one round of voting in Werewolf.)

This means that, at least in theory, most of the gameplay should consist of interactions that the authors haven’t pre-scripted and that the players are bringing into the experience themselves — as in a good game of Diplomacy, what you remember isn’t the moves you wrote down, but the interpersonal finagling you had to do to get there.

It’s an idea reminiscent of things that we experimented with in the early multiplayer demos of Versu. There we found that there was a certain added intriguing quality to social interaction with characters when you knew that some of those characters were being handled by your friends rather than by character AI. Versu was only designed to let players pick actions and speech acts from a menu, though, whereas Velvet Sundown allows free dialogue and movement actions, thus placing the emphasis much more squarely on person-to-person improv and less on social simulation.

Some of the YouTube videos of people playing Velvet Sundown are pretty interesting, and showcase both the good and the bad aspects of this design. On the one hand, it’s true that surprising things can happen: some of the playthroughs involve characters teaming up romantically in unexpected ways, sometimes they have bizarre showdowns, sometimes they follow one another around or try to get each other alone in natural-seeming ways. Sometimes they seem to have a really good time just using the game’s taser on one another (the use of which causes another character to crumple and become inactive for a moment).

It also looked as though the characters would automatically perform angry gestures if certain words turned up in their dialogue, which is cool if true.

It’s not quite the interactive theater of my dreams, though. Some features of the game tend to push the players towards goofiness and away from naturalism. There’s text-to-speech available for what characters say to one another (so you’re never hearing another player’s real live voice), but that makes for some funny misreadings as well as a significant interaction delay, because your conversation partner has to type something and then the computer has to read it to you. (You also get to see the words they typed written out, but because of the reading aloud, there’s still a sort of pressure to delay and not get too far ahead of it.)

When I played myself, my experience was at the less-dramatic end of the spectrum. I found myself in a group that as far as I could tell consisted mostly of fourteen-year-olds who had already played the murder scenario many many times and had a streamlined procedure for resolving it.

This made for some pretty odd playing. For one thing, my comfort in roleplaying a supposedly flirtatious character was undercut by the strong suspicion that the other characters were controlled by underage players. In that context, even a very PG portrayal of my role still seemed super creepy, so I pretty much stuck to impassively telling people about the evidence I’d encountered.

My wooden performance turned out not to matter that much, though, as roleplaying wasn’t what anyone else was doing either. One character asked to talk to my character privately and then with a lot of winky faces asked what my Steam name was. Another character produced a lock of brown hair, gave it to me, and then logged out of the scenario entirely, leaving his NPC figure to stand motionless in the corner. Other characters quickly arrived at their own conclusions about the murderer and went off to sign the logbook pretty early on; one seemed mildly exasperated that I was taking so long. It all seemed to accord with some set of social rules and obligations about how to play the game which a) everyone else there knew but b) had little to do with the game as designed.

I imagine it might be a very different experience to play with friends, and the YouTube videos give some idea of what this might be like. But if you’re inclined to try it out yourself — and I do think this is an interesting idea whatever its current state of development — I’d recommend trying to round up a posse of known friends who want to play it in earnest. (Given the time-constraints of the commission and the need to do this on a specific weekend, I wasn’t really in a position to try that myself.)

In any case, this was quite an interesting design space to think about, because it really straddles the line between conventional IF design (planting props and information for the players’ first experience of the story) and RPG design (coming up with mechanics to allow characters to resolve conflicts, but in a pretty open-ended way so that those conflicts can arise however they like); plus some thinking about randomization and replayability, given that at least some portion of the community likes to revisit these scenarios a lot.

I won’t go into huge detail about the thoughts I had about that, since those are part of the deliverable that I gave to the team and might be a bit of a spoiler. But it seemed to me that there should be mechanics that created both need and risk: needs being things you required from other characters, and risk being reasons why there was something at stake in communicating with them. Furthermore, it’s difficult to do strict limited information mechanics when the characters are able to talk to one another freely about whatever they want, without that dialogue having any mechanical effect. It really carries its own set of challenges, and I’m curious to see where the project will go from here.

(Disclosures: this blog post was written with permission from the Velvet Sundown team. It concerns about an experience I had using a premium account that was provided to me at no charge. The blog post itself was not paid for by them; neither the time I spent playing nor the time I spent writing this blog post counted in the three hours covered by the auction.)

3 thoughts on “Three Hours with Velvet Sundown

  1. Pingback: Multiplayer Twine: Bluedorn (Ryan Veeder), Salvanas (Andrew Plotkin) | Emily Short's Interactive Storytelling

  2. Pingback: Writing for Seltani: an Aspel Post-mortem | Emily Short's Interactive Storytelling

  3. Pingback: Chatbots as Narrative Platform | Emily Short's Interactive Storytelling

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s