The Big House - Scenes as Currency Structure

Lately I've been reading Burning Empires and playing some board games and I think I've hit upon a system for structuring the game, something lacking in my earlier design thoughts. I love the idea of using a limited number of scenes as currency that must be expended to achieve your goals in Burning Empires, but I wanted something to better capture the unpredictability and sudden violence of prison life.

So I think play will be organized into Cycles, during which each player can pursue a goal through 4 scenes. Extra successes gained during conflict resolution can be saved and rolled over into later scenes in a Cycle (encouraging planning), but after a Cycle ends, all left over successes are lost. The GM would get 2 scenes for each major opposition faction and several wild scenes that can be used by any NPC.

So there are 8 types of scenes, one focusing on each of the traits. So if you want to kill someone, you can only do it in a Shiv scene, while if you want to score some drugs you need a Yard scene. Your character's intent must correspond to the type of scene you chose to have. At the beginning of each Cycle, 8 cards (one for each scene type) are shuffled and then dealt out in a row, the first four face up and the last four face down. The cards dictate the order in which the types of scenes will be played out, and, of course, you're not sure what order the last four will come up in.

Once everyone sees the order the scenes will be played in, each player secretly writes down the four different types of scene they want their character to have this Cycle and their goal for the Cycle. The GM does the same. You cannot have two of the same type of scene, although you can be dragged into a scene type again if another character targets you with it. Then play begins with the first type of scene, starting with one of the GM's NPCs and then going to a player and then alternating between the GM and other players until everyone who wanted that scene type has had their scene. The scene types are played out one at a time until the Cycle is finished and each character has achieve or failed in his goal for the Cycle.

Example: Joe wants his character Vern to finally kill his nemesis Beecher. The scene cards are dealt out in order as follows: Yard, Snitch, Guilt, Block, and four face down cards. Joe devises Vern's plan. Vern will have a Yard scene to procure a weapon, a Snitch scene to get the guards to turn a blind eye in exchange for ratting out another inmate, a Block scene to get his Aryan Brotherhood buddies to help out in the killing, and finally a Shiv scene to attempt the murder. Vern hopes to roll extra successes in the early scenes to help in the Shiv scene and put that bastard Beecher out of commission for good. Of course, who knows what Beecher is planning...

So my open questions:

1) The planning and goal creation parts of this structure need mechanical support to have this scene structure make sense. I need to ensure that the characters all have strong story goals that can broken down into Cycle-sized pieces, kind of like BE's Infection mechanics but much more flexible.

2) If the face down cards don't come up like you want them to, what would be an appropriate consequence - Forced to follow through anyway? Able to abort and lose one (or all the rest) of your scenes?

3) Is the whole thing too mechanical (betraying its boardgame roots), and does it capture the kind of play that I want?

As I move forward, I'll see how each answer emerges. In the meantime, feel free to chime in with any thoughts.