Inspired by: this. Related to: this.

Rashomon, at the core of the movie, is about the subjective point of view of reality. Four people see the same events transpire. While the events are roughly consistent in the facts of the events, the framing and motivation attached to each of the events differs.

The "play it straight" setup of the characters recounting their experiences is out, the low-bandwith voice and hand-waving in lieu of direct experiences will make it difficult for the characters to frame their stories with any level of different subjectivity. As the players listen to each other's stories, the subjective consensus tends to converge on an accepted reality.

Instead of going pure rashomon, it may be preferable to go with something like Leverage's Rashomon Job. While not quite as "hard hitting" as the original movie, it contains more action and is resolved far more neatly.

What system supports and allows players to highlight characters with different subjective constructions of reality? What instructions and encounter setup will be necessary to provide for the cognitive dissonance that provides enjoyment in the movie?

  • 3
    \$\begingroup\$ "As the players listen to each other's stories, the subjective consensus tends to converge on an accepted reality." It may be possible to work around this by framing the play in a different genre. RPG tends to use journalistic or realist novel modes. Boasting sessions encourage participants to accept, extend and elaborate past contributions; for a modern example, the rap battle. Shifting from a naïve representational mode of literature, to a competitive mode (at the literary level) may help. "Well he said he gone done murdered me but I playin' possum, haw." \$\endgroup\$ – Samuel Russell May 15 '12 at 22:57

I'm not sure there's a system that particularly supports this, but the scenario setup that immediately leaps to mind is providing backstory to the players individually before the game.

Tell the group that time is a concern, so in the day or so before the game you're going to individually contact each one to explain the backstory that leads up to the game; that way the group can jump right into the action as soon as everyone is there. Use these individual briefings to feed each player their different perception of the events that occurred. DO NOT let them know that you've given each one of them a different description: let them figure that out on their own once they start playing.

Type up each person's backstory description beforehand; even over the phone, people can tell when you're just reading off a script, and it won't occur to them that you made a separate one for each of them. If these briefings are in person, do not give them the typed up description: you want them to be comparing their recollections of the description, not the typed up descriptions themselves.


I actually pulled something similar off at a convention once as a tournament. Six characters per table, three rounds (18 tables => 6 tables => 2 tables). Myself and my co-author let the different GM styles from the previous rounds provide the confusion in higher rounds, but worked HARD to institute the proper feel in the first round. Here's what we did.

We were playing a white-wolf hunters game where the characters did not know, initially, that they were fighting supernaturals. We touted it as "White Wolf Mortals". Many of the NPC's were Werewolves and Vampires, but until near the end the characters never saw them outside their human forms (we were aiming for an X-Files feel). We had two agents from each of three agencies. Each character packet approached 20 pages and included personal information, dossiers on the other characters in the party (written differently, and sometimes with conflicting info by the different agencies and, less so, for the different characters), dossiers on the major players (again, written differently), and autopsy reports on the dead bodies that were the "hook" for the game (again ... yeah).

We enforced a strict 10 minute time limit for "familiarizing yourself" and dove right into the action, leaving them to scramble to cover all the information in their packets. Myself an my co-author watched the first couple rounds before running the finals ourselves. It was one of my finest moments as a storyteller as I watched someone shout "Wait ... it didn't say that in MY dossier" and dive across the table to snatch someone else's packet. And again, during the first break, to watch the players walk off with their packets muttering about being confused, but being convinced the answers were there in the packets ... complaining that all that was missing was a cigarette smoking man standing there to be condescending to them.

I would say that there are two things to make this successful in the short term: LOTS of prep work and buy in from your players. Even so, I'm not sure it's sustainable. This game style requires intense work on the part of the person running the game, but can be incredibly rewarding to watch it unfdold.

  • \$\begingroup\$ I assume you meant 'X-files' and not 'X-men?' \$\endgroup\$ – GMJoe May 16 '12 at 4:27

This reads like a pretty sweet game-design challenge--I can't think of any games that do specifically this (Lies of Passage, one of this year's Gamechef finalists, might be a near miss, however the lying is collaborative there) but I imagine that the key to preserving cognitive dissonance would be to play through multiple subjective scenarios that still had the potential to be/contribute to the 'real' scenario, and to have some mechanism that resolved the actual order of events from those building blocks. Make that 'convergent accepted reality' an explicit part of the game. That way each side of the story, while not totally true, would still have high stakes because it would affect the actual resolution. You could do something like play through 3 different backstories to an event, and then have the narrative results of, say, each critical success and critical failure rolled in each scenario be compiled to form the true events.

Now that I think about it, Dirty Secrets could be made to serve this role, though it wasn't explicitly designed for it. It's a noir crime game where the actual perpetrator isn't pre-determined and emerges over the course of play. As such, someone playing a character could maintain their innocence of a crime with perfect sincerity, only to discover that they dunnit after all. The game requires some present-tense action by an investigator, but it would easily support the main action being a judge or detective listening to various characters' version of events and trying to synthesize the truth.

  • \$\begingroup\$ I would love to see some "determine reality" mechanics, maybe built on top of donjon or minimus? \$\endgroup\$ – Brian Ballsun-Stanton May 15 '12 at 18:28
  • \$\begingroup\$ @BrianBallsun-Stanton That seems like it would work really well, particularly with donjon (haven't played either game, but did some googling). I kind of want try it out now. \$\endgroup\$ – RSid May 15 '12 at 20:39
  • \$\begingroup\$ Drop by chat and we'll work out the rules and post them here. \$\endgroup\$ – Brian Ballsun-Stanton May 15 '12 at 22:03

I would start with Fiasco and make a custom playset if none of the published ones look relevant. Players take turns doing freeform scene building which closely matches the subjective nature of Rashomon. The focus on character relationships and needs is also excellent game fodder. Finally, the fact that it only does one-shot games means things get wrapped up before they get too confusing.


Not the answer you're looking for? Browse other questions tagged or ask your own question.