It worked really well, actually. Better than the House of Masks playtest went. We had a fun game, and I got some good user feedback and good ideas on how to move forward with the game.
We had a total of ten players, including myself. This is far too large for a single game, so we split into two groups of five. I played in one table (Table One), so I don't have as much information about the other game (Table Alpha). It seemed like they had a successful game, albeit slightly less absurd than ours from what I saw. Well, Hitler turned out to be a shapeshifting Yak hiding out in Tibet in the 1960s, so it still was pretty ridiculous.
Department Nine is a very odd sort of collision of several genres. I used to describe the genre as "Time travel/espionage/classical Greek Tragedy/comedy". That leaves off the white collar/bureaucratic absurdity, though. It's like watching Minority Report, Back to the Future, The Office, Burn After Reading and Oedipus Rex all at once, perhaps. We had a were-swan manager being blackmailed by his employee, who was channelling advanced technology (eventually, a doomsday device) back in time to Benjamin Franklin to ensure the Yanks won the Revolution. We had an illicit office romance involving going back in time to have megafauna safaris. We had glyptodon races interrupted by sabertoothed tigers and flying car police chases interrupted by monkeys flinging poo. We had a spy disguised as a pizza delivery guy get in a three way sex scene with a woman and her future self. I don't even know all the details of the other game's stuff. As I said, it combined a lot of different genres into one hilarious mess.
The initial explaining the rules was a bit rocky. This was probably my fault, due to a lack of sleep the night before. I managed to start the game without explaining Flags, which are one of the most central mechanics in the game. I also forgot about the Fate Manipulation rules until Adam reminded me about them, but it became clear that those were a vestigial rule no longer needed, and they were duly ignored throughout the actual game (it wasn't a great rule even in previous versions of the game).
The new randomized setup phase managed to generate some wildly bizarre and silly situations. Said weirdass situations did conform to the mishmash of genres mentioned up above, so the random cards as genre creation tool seemed to work very well.
After the fact, it was discussed and determined that the random setup is too random, and it was proposed that the player have a bit more control. As the game stands now, the player gets cards dealt to them entirely randomly to determine their relationships, fate and what their scenes involve. Amber suggested giving the player all six cards at once and letting them decide where they will fall. We tried out this alternate setup after the game, and it felt like it would work better. I think that will address many of the issues that came up over the course of the game in trying to fit together weird and incompatible bits of the game.
It was also suggested that there should be more than six categories: as it stands, the game has you script one element in each of six categories (Time Travel, Espionage, Bureaucracy, Greek Mythology, Family and Romance). These six led to some overlap. I think future versions will have more categories of which the player picks six. Ideally, there would be ten or so categories, but I'll have to think about what additional categories would fit into the game.
The new sortition conflict system worked amazingly well. It adds a bit of randomness, which is good. And even better than that, it adds more detail to what is going on in a conflict and keeps everyone at the table involved. If you're sitting to the side in a conflict, you still can add details and influence play and such, which is a lot of fun. Several times in our game, two players were pushing the conflict in different ways, and then a third player would pay a token to place a third color of bead into the bag, surprising everyone. At that point, the conflict went off in a new, surprisingly and entertaining direction.
We had Pete try a conflict against himself at one point, which is a weird enough thing that it needs to happen in playtest at least once. I think he was trying to earn the "NPCs win conflicts" flag. It seemed like it worked just fine, though. Like Pete was just trying to use the bag as a randomizer to decide an outcome for himself.
The endings phase was the new rule I was least certain about, but it seemed to do its job without major problems. Having other players decide your success or failure is a major reversal from previous versions. But it does make everyone more involved in each other's scenes, and removed the necessity of the Cause scenes from prior versions. Ross suggested the excellent rule that "whoever donates the most chips to a player gets to narrate their ending". This gives some small incentive to give big payouts to certain players rather than spreading the tokens evenly around to give everyone a happy ending.
The difficulty of the ending phase was set at five chips: you needed a total of five tokens donated to you for you to have a happy ending. This seemed like about the right value, as we had several happy endings and several unhappy endings in each game. In the text, the ending difficulty is tied to the number of players and to the defunct Fate Manipulation rule. So in both games the difficulty would have been 5, but in a smaller game it might be different. We need to try a three player playtest at some point and see how well that works. In a smaller group, I think that we'd see the Flags spiral in value more slowly, so maybe three is the right difficulty. Or maybe not, I don't know. It may depend on your group.
In the Table Alpha game, they had two PCs decide to accept their fate and end their game early. This was unexpected, but may need a sidebar or something to cover this situation. Mostly, it worked out fine anyway, as we just decided that the other players shouldn't give those players any chips in the final end scene.
There's definitely a possibility of someone sitting back and not exerting themselves. I didn't have any chips for a lot of the game because I rarely introduce Flag elements myself (I wasn't on the top of my roleplaying game, exactly, though I had a lot of fun and bizarre adventures). In a somewhat different matter, Chuck in the other game apparently got a lot less face time than other PCs. As it stands, the game has no guarantee that anyone will get equal facetime. Each player frames three scenes, though, so they can get more facetime for their PC if they really want. I don't know if that is enough, or if we need another balancing tool of some sort to reward giving more screentime to reclusive players. Or maybe it is okay that Chuck decided to primarily play NPCs and frame scenes for other characters and act as a sort of pseudo-GM role. (Actually, in those terms, perhaps this aspect needs explored further.)
Final note: once again, no characters had names. I seriously need a damn rule making sure every PC gets a name, because it is forgotten about every time that we play.