Mechanic. That's key, right? Got to think about the mechanic. MUST think about the mechanic. It's all about relevance-- it's got to fit the game, enhance the premise, or there's friction. They might both be leather, but a glove makes a lousy shoe, or some such metaphor.

So, I have a nagging fear I may be doing my premise a disservice when, even after some examination of the story, I still find that I like the same garden-variety "dice pool" mechanics that I did back when I was saying "Screw it-- I'm using garden-variety dice-pool resolution mechanics!" I don't want to be systemically revolutionary, but I don't want to short-change myself, either.

First, a recap of the premise*:

The game puts players at an ordinary day at the office, although there is (recommended to be) some "amnesiac" start to the game (a cliche I picked up from the Interactive Fiction world). It's rather humdrum until a truck-bomb explodes outside the building, and characters have to find their way out alive, and otherwise deal with the situation. The extraordinary but still rather pedestrian adventure gains a twist when, after a preordained amount of time, the scene resets (in-game, characters "black out" mid-action) and the characters find themselves back at the beginning.

The story consists of two parts: The backstory and forestory. The forestory is the repeating situation I'd mentioned, and is the part of the story overtly revealed to the players/characters. The backstory is that this event has already happened some time in the "past", and that the whole repetitive event is taking place synchronously with the anniversary memorial dedication of the site. (GMs are encouraged to include subtle environmental cues that suggest this.) All of the players' characters-- except one-- survived the event. This person must die in a replay, either through accident or outright murder, in order to release the rest. The main player task of the game consists of trying to identify this "Fated" person and their dealing with that fate.

The game does have a few story-specific mechanical elements--

The most notable variation is Deja Vu cards. The cards are written by the GM pre-game, and consist of snippets of memories from the "actual" event. Cards are compiled pre-game into coherent numbered decks by the GM, with higher numbers containing more revealing memories. At the start of the game, players choose their deck. This also determines which player is the Fated. The game is played using d10s (Keep the math metric, y'know?), and there are a total of 11 numbered Deja Vu cards for each player, as well as one standard common card, "Card 0", which consists of the opening phrase "You are at the office building where you work. One of you will not make it out alive." (or some such post-edited revision thereof.)

Cards are kept hidden from the players until players roll to reveal them. At the end of each repetition, the players roll a d10 for a Deja Vu card. 1-10 gets that numbered card, unless that player already has it, in which case they receive no flash of insight that round. If the player rolls a 10, they can roll again for the "11" card. If they then roll a 1 or a 10, they take the 11 card as well. The "11" card should tell, in no uncertain terms, whether the player is the Fated. Players do not have to reveal these memories to others (and should do so in-character).

The recycling nature of the game changes some of the common rules regarding health and death, as well. I'm planning on using a "scale" based health mechanism with various impairments at certain levels. All player-character health does return once the cycle resets, but if a character dies, they die permanently. The game-world revolves around the player-characters, and NPCs' fates also rest on their interaction with the PCs. If an NPC has a significant enough dialogue with a player-character to become more than just a piece of the scenery (at the GM's discretion), they become "vulnerable" and will also die permanently, and be missing the next go 'round. This was started as an answer to the question of "What about NPCs?" Without NPC permadeath, the game would be far less lush, in that the characters would be simply fighting to get themselves out of a two-dimensional stage play. With all-inclusive NPC permadeath, however, players have the added element of personal heroics, but by the third or fourth explosion, the office would be mostly empty. Conditional NPC permadeath, I think, will be a compromise that adds a number of possible interesting angles to the story.

Still, though, a lot of the moment-to-moment conflict resolution mechanics I'm considering are pretty well straight-up-the-middle. D10 dice pool based on skill set, success/fail with threshold and degrees, with an emphasis on GM rule for simple tasks. (Y'er basic modified-WW rollup.) My mechanical desires are something that has the granularity to accurately simulate realistic situations, but, within that requirement, runs as simply and unobtrusively as possible. Simple math, obvious answers, and no lookup tables. Weapon use will probably be minimal enough to roll it into a general "tool" modifier, considering the limited environment, although I may, upon reflection, have to work some more meat into environmental damage, as that will be a heavily contributing factor.

Really, I'm just wondering (oh experienced audience) whether this style of resolution mechanic would lend itself well to the game described, or if I'm overlooking pitfalls or missing opportunities.

Background, for those interested:
Premise (HTML)
Power 19 (HTML)
Manual Snapshot (the half-written draft) (PDF)

* Noob question: is it a good idea to repost a simple game synopsis, at least until a game gets more than one topic above the fold :), or is it assumed that the reader should search the board for previously posted information?

At the end of each repetition, the players roll a d10 for a Deja Vu card. 1-10 gets that numbered card, unless that player already has it, in which case they receive no flash of insight that round. If the player rolls a 10, they can roll again for the "11" card. If they then roll a 1 or a 10, they take the 11 card as well. The "11" card should tell, in no uncertain terms, whether the player is the Fated. Players do not have to reveal these memories to others (and should do so in-character).

The most notable variation is Deja Vu cards. The cards are written by the GM pre-game, and consist of snippets of memories from the "actual" event. Cards are compiled pre-game into coherent numbered decks by the GM, with higher numbers containing more revealing memories. At the start of the game, players choose their deck. This also determines which player is the Fated.

Still, though, a lot of the moment-to-moment conflict resolution mechanics I'm considering are pretty well straight-up-the-middle. D10 dice pool based on skill set, success/fail with threshold and degrees, with an emphasis on GM rule for simple tasks. (Y'er basic modified-WW rollup.) My mechanical desires are something that has the granularity to accurately simulate realistic situations, but, within that requirement, runs as simply and unobtrusively as possible. Simple math, obvious answers, and no lookup tables. Weapon use will probably be minimal enough to roll it into a general "tool" modifier, considering the limited environment, although I may, upon reflection, have to work some more meat into environmental damage, as that will be a heavily contributing factor.

Good points, all... this is why I keep comin' back.

I'd intended to keep the cards rare. Part of the situation is puzzling out the roles, and I'm seeing the cards being more of an "extra bonus hint" than a dependable information feed. The nice thing is that with there is a natural decay. Everyone will receive one card the first round, and chances are good for the next few rounds that people will get something. I suppose the effect and effectiveness of this system will be something to be eagle-eyeing during playtesting.

I disagree about pre-play burden being too terrible. I'm figuring that there's a sentence or two per card, just a quip, and considering that the players are intended to be socially familiar, and because the cards aren't publicly shown, many cards can be the same facts from different perspectives, or the same outright. It might throw in a roadblock for "hey let's play" immediacy, but with any prep time at all, it's surmountable. (The GM could even just write up a set of simple synopses, and write up the specific "cards" as needed, I suppose.)

I've been mulling over the comment about it being a tactical game for quite some time now. (This reply has been living, in various stages of growth, in a text file on my laptop for a good three days. It's not that I'm that slow a thinker-- I just keep being interrupted when I go to type.) I do think one of the weaker points of the idea thus far is that I'm not sure what the overall "win" or goal of the game would be, and from there how to reward or encourage certain styles of play. My thinking is that with the plot so constricted, I don't want to further constrict the possibilities by setting down ordained values in the system-- hence my inclination to go with a generic mechanics system that would control little more than immediate pass/fail on specific events.

After this reply, though, I have started to think about what the game goals are. In the least, the game is a simple "get out alive" scenario. It is, however, divided into two portions-- the first being to gather information and determine everyone's roles and state. This portion, along with the necessary self-preservation, is rather tactical. Once the situation becomes clearer, the question of how to deal with it comes into play, with the internal and interpersonal issues coming into play.

I have to agree with your statement (or, rather, your leading question) about the non-necessity of accurate simulation. I suppose that accurate simulation is more of a personal preference/hangup of my own than something that is necessary to something like this game. I'm still thinking that a generic resolution mechanic is the way to go (barring further discussions/personal revelations), but I'm planning on simplifying things down and lightening it up a lot.

Cool, FLEB.

I'll be thinking about your game a lot in the next while.

I have a game I'm running in the upcoming week called Cheap, which is all about figuring out how to work people and interact with them in this place, throughout a non-linear and fairly disconnected story progression. Basic pitch here: http://raos.naos.org/smf/index.php?topic=38.0 (note that that forum is a strictly Vancouver-based one.)

Anyways, if you see the parallel that I do between Cheap and the "constantly resorting how to deal with people in this nonlinear tale" aspect that I see in Glass & Steel... let me know if you want to see some AP, or possibly video recording.

Yeah, I'd definitely like to see anything you've got (or end up with) handy. It sounds like a really interesting idea, although I confess I haven't seen half of the movies in the prereq list.