Secrets from the Floating City
Monday, September 5, 2011 at 10:20AM
Once upon a time, a friend put me in touch with some people who were looking for some help with a project. Within days, I found myself on Skype talking to Thomas Dolby, man, can you believe it? and learning about one of the most richly embroidered and charming story worlds I have ever had the privilege to work with. It was meant to be a small game, mostly text; and there wasn't much budget for it. But I was very much enchanted by the idea -- and to be fair, not terribly occupied at the time -- so I signed on for a token fee, just to be a part of it.
That was in October of 2010. I had no idea what I was in for. It has been an adventure.
In the months between, I launched half a dozen other games, and sometimes Floating City got short shrift as a result. During the game's actual run, I wasn't able to do active and ongoing moderation and felt a lot of guilt over it; but the stuff that pays your mortgage and feeds your children does have to come first. Still, Floating City was a wonder, a marvel, and in some ways even a homecoming for me, and I am profoundly grateful to the players and to Thomas Dolby himself for letting me be a part of it.
The Game
For those of you who aren't aware, Floating City was basically a trading and set-collection game, in which the sets you were collecting were items from Thomas Dolby songs. Players were assigned to one of nine tribes on signup, and had to work their way to the North Pole through trading those items. Collecting all of the items associated with a particular song would give you a free download of that song.
This was all interwoven with a fiction played out through an online newspaper, forum posts, and chatrooms; some of those inventory items were special intelligence items, shedding light on the backstory that led to the players' mysterious situation. Periodically, there would be in-game events where a map location would briefly become available, and players could solve a simple challenge in order to get new items and information. The newspaper would sometimes cover oncoming hazards, from time to time -- players would have to make and submit a patent in time describing inventions to protect them from, say, marauding pirates, or risk losing items and progress.
There was prizing, too, and achievements; players garnered reputation from trading, posting on forums, sharing on social media, and so on. Milestones of "reputation," as the points system was called, would earn you more items and currency to trade.
And this is... this is my very brief description of it. To think we set out to make a simple game!
Development Process
I could write a very long and tedious post about various design decisions and why we made them: Items weren't distributed randomly, but instead were weighted so we had some control over how many popped into the world. Some items were only available to a single tribe, with the hope that each tribe would have a unique trading advantage, hopefully offsetting the disadvantage of being in a small tribe to start with. Trading was your sole source of movement, and you had to rely on someone else for your own forward progress, to force players to work together and plant the seeds of larger-scale cooperation.
The only element anyone cares about is scoring, though. The scoring mechanism caused... some significant controversy, both when it was secret, and after it was revealed. So I'd like to talk a little about why it was the way it was.
The grand prize for Floating City was always meant to be a private Thomas Dolby concert, and for reasons of art, Thomas wanted nine tribes. This made a lot of design decisions for us: the game had to be competitive because we had to have a way for someone to win a prize; we had to do geosorting so that we could determine where to hold the final concert.
It also meant that we'd probably be dealing with tribes of wildly varying size. But we weren't sure how much they'd vary, or how those tribes would shake out, exactly. We evened them out as much as we could by looking at the physical distribution of Thomas's existing fan base, but even so, there were sure to be wild disparities. We needed to strike a lot of balances so as not to give either end of the size spectrum an unfair advantage.
I've described the precise formula in detail elsewhere -- it amounted to an average of the scores of the top 20 players in a tribe plus the whole tribe's average distance from shore multiplied by 100. That was the result of walking a lot of fine lines and asking ourselves a lot of questions and thinking about a lot of edge cases.
Would it be fair for a few highly involved players to be dragged down by a tribe in which a lot of players signed up, played three days, and then left, leaving abandoned ships behind? Would it be fair for two or three very active players from a small tribe to earn more points than thirty, who had less reputation but were more cooperative? We felt that average distance was a great metric for our score, on the assumption that all tribes would have roughly similar ratios of abandoned accounts; but rewarding achievement, too... that was harder.
We could have gone by a sheer percentage; but then a theoretical one or two manic players might win the game for a tiny tribe, and we wanted collective achievement to be the point, not individual achievement. Compounding that, we had literally no idea just how many players we'd be seeing. A hundred? A thousand? Fifty thousand? Averaging your top five percent (or your top thirty percent) wind up being very different metrics, depending on if your player base winds up being a hundred players vs. fifty thousand.
We settled for something like a per-tribe leaderboard -- that top 20 players -- hoping that our smaller tribes would get enough active players for this to be achievable for them. This meant that a winning tribe would need a critical mass of active players making progress, and not a mere handful. Cooperation and achievement together, even at the highest levels of the game: Bingo.
Ultimately, the scoring worked far better than I had dared to hope. By the end, both the top-20-players measure and the average-distance measure were of equal weight, meaning that we were, in fact, rewarding both achievement of the most active players and communal progress of the whole tribe evenly.
On the other hand, small tribes never had the fighting chance I'd hoped they would. Luckily, this was readily solved by our alliance mechanic, which applied the scoring formula to the pool of players from the combined tribes. All said and done, I'm content that it worked the way we meant it to.
Personal Accomplishments
Floating City held a lot of firsts for me.
For one thing, this was the first time I was personally denounced in a game I'd made (that link goes to a locked forum -- sorry, it's there for my own reminiscence). I don't believe anyone knew my role at the time. A curious milestone, that.
The person denouncing me had sent me a private message essentially asking to speak to my manager at one point, another thing that had never happened to me before -- at least, not since I was a receptionist at a computer repair shop too many years ago, trying to tell an angry man that his laptop was out of warranty.
The first time I ever sprained my wrist. I took a bad fall off a tall pair of shoes in early August, while traveling for my aunt's memorial (another very unhappy milestone that happened during the game), and have been in a splint ever since. Even still. And me with a book to write.
On a lighter note, this was my first time working directly with someone you could call a celebrity! (For America 2049 and for 2012, I wasn't directly involved in any of the filming.) I am delighted to assure you that Thomas Dolby is a charming, kind, thoughtful man, and not anything like you imagine a celebrity might be based on tabloids. Completely sane in all the ways that matter. I'd love to work with him again one day, in fact.
As was pointed out to me by a player, this is the first time I've been so very accessible to my player base while a game was running. The first time I was speaking as myself, and not as an entirely fictional construct. I really loved being able to do that. I'm going to have to find a way to do that again soon.
Finally -- Floating City reminded me just how much I love seat-of-the-pants, performative storytelling, even despite the scrambling and drama. It wasn't a perfect game, but for all its flaws, I think it's one of the ones I've wound up loving the most.
And now: I'm happy to engage in a longer discussion of Floating City from a design perspective over here, both with players and with other designers curious about what we did and how it all worked. If you have a question you'd like to ask, go on and ask it; if you have a point you'd like to make, go on and make it. Pull up a chair, pop some popcorn, and let's have some fun, shall we?
Reader Comments (9)
Hi Andrea!
Doggedly Jo, here
Thanks so much for your post, and thanks so much for the game.
It was interesting to see that link to the hobvias shirt drama in there, I was wondering if you'd mention that. I'm glad it will be a part of the history of this game, though I am so sorry it ended up being an unhappy milestone for you.
You know how much the game meant to all of us, and how much we all enjoyed it, and I hope you are coming away with that accomplishment.
Lots of love from Angelica
It was certainly a pleasure playing/working with you! In my opinion, the game was a resounding success and absolutely fascinating to see how human nature acted and reacted within such an intense environment dislocated from reality on so many levels! It was also fun to play ;-)
Hi, Doggedly Jo! nice to see you here. ^_^
The hobvias thing came at a pretty awful time for me, personally, since my aunt had passed away over the 4th of July weekend, and so of course I was blue over that, and trying to get SXSW proposals together, and arranging book contracts and meetings with my editor, and and and it was so much right then. I was tired and overwhelmed, and then really, really surprised at how that played out.
But it's a thing that happened, and if you're not honest about what went right and also *wrong* in a project, then you don't get better, do you?
Nikki, it was likewise a pleasure working with you. All of my most favorite projects can also be described as "an interesting social experiment." ~_^
I have a few routes of discussion I'd like to explore, if you don't mind.
1. After the game was over, I was discussing the game with another player. I was disappointed by the competitive nature of the game. I linked her to the keynote address that Elan Lee and Sean Stewart gave at ARGFest 2007 in SF. One of the things they mentioned was that their initial intention with the Beast was to have the players divide up into teams and compete, but that after seeing how the Cloudmakers worked TOGETHER so amazingly, they gave that up as, "so last century." Sean even mentions the collaborative process that occurred almost restoring his faith in humanity.
Considering how paranoid and upset and cut-throat and to be honest, kind of horrible people acted toward one another in several cases, would you possibly rethink going the competitive route? Can you imagine how different the experience you had would have been if they'd gone that route with the Beast?
2. Did you regard what I was doing with the dolls as "game-jacking" at all? I knew that creating my own story didn't actually affect the in-game story, but at least one person thought I might be in-game. I tried to let them know I wasn't but who knows when the lines might get blurred *cough*Dr. Argleton*cough*.
3. I think the scoring system was flawed. It generally ended up giving the lead to the largest team(s). There was some slight deviation, but teams with 2-3x the players of another team would always be in the lead. period. I am curious to know some details (like how many voted in the Book of History challenge), because I'd really like to figure out how it all ended up working statistically. When I was doing some analyses on it, I was using a lot of guesses.
I think averaging would have been better. Though it might have given an advantage to a smaller team, that could have been mitigated with bonus rep for rep earned by people you invited. I actually thought of this immediately after the game, and now I recognize a form of it in practice in the Reveur rank for the Night Circus. Also, I kept thinking there'd be a challenge of some sort where our teams would have to work together to get something from the pole to some other boat near the coast. That's another way there could have been a reward for a large, well-organized team.
I mean, I do realize that hindsight is 20/20. And I'm not suggesting that the system was inherently unfair like some have. And you did provide a way for smaller teams to form alliances to make it better. I'm just trying to discuss ways that might be better for next time.
You ask some excellent questions.
1. It should be clear from my past portfolio that I prefer making collaborative or at least single-player experiences to competitive ones. That said, I don't think there's anything inherently wrong with a competitive model... team spirit and competition can drive some really interesting behaviors, and friendly rivalry can be a lot of fun, too. The problem is with prizing. It's intensely difficult to make an internet-based game that feels fair, is legal, and gives out prizes.
If I had been designing basically this game with fewer starting parameters, then I probably would have removed the top prize, which would have had huge cascading implications for everything else, even if I kept the team paradigm. Or, if I'd had unlimited resources to throw at the game, I would have moderated for tone much harder from very early on, to try to cut down on some of the hostility before it became too deep-rooted.
2. I didn't consider what you were doing to be game-jacking at all. In fact, what you were doing was exactly the sort of thing we had most hoped to see, you were actively helping to build out the world. ^_^
I hear you on the not-clear-distinction between players and GODs. I wish I could say I was clever, and the blurry line between players and creators was porous on purpose... but actually that was an artifact of a situation where a lot of the collaborators on some elements weren't active parts of the design team on the mechanical and tech side. There were gradations of GODs, from people who basically made the game run under the hood (like Captain Smith) to more peripheral friends and volunteers who didn't have any better an idea what was going on than any other player.
I'm not sure if or how I'd change that, knowing how everything worked behind the scenes.
3. I'll agree it wasn't the platonic ideal of scoring, so let's get that on the table right away. ^_^
The assumption that a bigger tribe would always do better is assuming roughly equal amounts of engagement. Which is... a pretty big base assumption. When we were making up the scoring scheme, we weren't even sure any of the tribes would have more than four or five really active players at all! So we were looking for a way that would reward not just a few active players, but active players encouraging activity from other players, too.
But let's take a harder look at an averaged percentage: If we'd gone with something like the average of the top 10%, that might have felt more fair all around. But it was a hard call to make when we didn't know if 10% was going to mean one dude... or fifty. And in the case of a tribe of ten people vs. a tribe of a thousand, it would've meant that the bigger tribe would have had an insurmountable burden of dead weight to carry.
If we'd done an average of all of the players' reps in a tribe... then every tribe would be hostage to abandoned accounts, and the deciding factor would've been who lucked into the lowest rate of abandoned ships. Which might have been perfectly even... but it might not've been, either. And since we already had distance potentially weighing a tribe down, I really wanted a way for just the most active players to be able to bring it up.
If I'd had more time to plan and more resources to model the math ahead of time, and to be completely frank if I'd even thought about it *that much*, I might've gone with an average of all scores for players who had logged in at any point in the previous two weeks. I wonder if that would've looked even more confusing than what you saw already?
As for sharing the exact poll results.. maybe one day if you get me in the same room as you I'll show you my Excel spreadsheet for that. ^_^ I... probably shouldn't put it out there on the internet, though.
This is one that I badly badly badly wanted to play but was so busy on my own projects. Sigh.
One of the reasons was the artistry of the website...it was so mad beautifully done and the level of detail and magic and intrigue, it really whet my appetite asthis kind of stuff is so important to me as a player/creator.
Thanks for sharing your production experience.
@Andrhia I had a few spreadsheets that I put together to try to figure things out.
Some of the ways I'd looked at it included:
A. Using a cutoff value for the low end of the scale. I think one of the guys from Frozen River tried a sliding cutoff that scaled with the progress of the game. That seemed like it had a lot of promise. assuming a certain minimum number of trades x 30 pts/trade + some number of posts/day x 10 pts/post. and then raise the bar that number of points ever day. The double-edged sword of cutoffs is that while they do get rid of the people who stop playing, they also get freeze out the new players.
B. Taking each players score and scaling it by their progress toward the pole.
thinking about it a little more, who should win:
Team 1 has one player who scores 1000 points.
Team 2 has ten players and they each score 999 points.
OR
Team 2 has ten players, and three of those score 1000 and the other seven get 900 points.
OR
Team 2 has two players, and though one scores 1001 points, the other only gets 900 points.
in each case, Team 2's average is lower than Team 1's.
I think the first case could go either way. probably leaning toward Team 1. In the next case, I'm definitely leaning toward Team 2. And probably think they should win in the third case.
it's not cut and dried. Even when it SEEMS simple (like a baseball game) you end up with people saying the only reason the Yankees are good is because they can buy better players. So maybe we could suggest that baseball scores be weighted by the amount the teams spend. ;)
Given the goals of the game, in your examples, I'd rather ten players with 999 points win over one with 1000; they've displayed collective achievement. The other cases are... harder.
The new-player thing is hard, too. That's another reason we couldn't have gone with an average of all reputation -- it would have actively penalized tribes who were good at recruitment, and in a very visible way.
As is, there was some penalty for recruitment -- but a short-term one, if those new players stayed active. We stopped pushing hard for tribes to recruit new players a few weeks before the end, when the window was closing for new players to plausibly make it close to the pole before endgame.
Thanks for the game experience and (as I see now!) the deep thought into the whole scoring mechanism.
My only regrets were probably technical ones on my own end - I was traveling quite a lot for both business and pleasure this summer and my internet access was quite limited. I think the game worked very well if you could spend a decent amount of time at it - not so well if you couldn't take the trades in close-to-real-time.
And I loved the dolls.