What must this scene do? Remember, this is in relation to your overall campaign. A combat scene does not need to invoke the combat rules of a system to be successful. This should be your core goal for the scene in question, the one thing that the scene must do in order to be successful. Simply put, if you can't check this off after a scene, it was a failure.
However, what that means varies from session to session and game to game. For example, in Session 1 of pretty much every game I run I throw in a combat. The goal of that combat is to let the players figure out how their characters work in combat. For that session the must of the session is the scene must make the players interact with the system mechanics through their character. That's it. Everything else is dressing.
The biggest thing about this question is that if you can't concisely answer what a scene must do, you likely need to re-evaluate the scene. It's ok if the answer is "the scene must fill a chunk of session to slow things down" provided that's not the answer for every scene.
Some other example key goals I have for various scenes in my head:
- The scene must show a different side of a prominent NPC
- The scene must introduce a plot important NPC
- The scene must show some of the danger here to the PCs
- The scene must establish some of the supernatural aspects of this area of the world
TL/DR: The must should be a short, concise singular objective thing as to what the scene needs to do in order to be successful for the game overall. It should not include things such as feels or impressions as you can not control the communication of those things (i.e. introduce new NPC vs. Introduce new NPC in a scary way.)
What should this scene do? Is a place for higher end goals in the planning. This is where you bring in ideas like themes and feels or moods. For me this is where I have my goals for execution. Must is what I need to do. It is the - for lack of a better term - mechanical core to how the story is being doled out. The core function of a combat could be to introduce a key NPC, or show a different side of an established NPC and those are both perfectly valid. The should is more in how I hope to convey this, or the path I hope to take.
As such, I like to think of my 'Shoulds' in connection with my key goal - or must - for the scene. For example, my Session 1 combat must make the players interact with the system mechanics via their character so they can see how they work in combat. However it also should communicate how I want combat to feel in this game whether that be "cinematic and exciting" for a super hero RPG or high fantasy, or more "gritty and violent" for sci-fi/dark fantasy/horror type settings.
The big thing on Shoulds is that you don't have full control over whether it is successful or not. You can try to make a combat cinematic and exciting, but if the Players aren't buying into it it may not happen. The other thing about Shoulds is they don't dictate if a scene was successful or not, but they do contribute to how good or bad a scene is. Again though, you don't have full control of that, how your players take things in is also a big factor.
For example, the 'must' for a scene could be to introduce the new boss for the arc the PCs are on. The 'should' in that case could be "
should be established as a real threat to the PCs." However, I can't really control that. If the Boss is introduced in a fight, and the dice aren't with them they're not going to be scary. If the PCs are just not pushed right, the boss may be respected without feeling threatened. Or they may be ridiculed and come across as weak or a joke.
Finally, you can have 1-3 shoulds for a scene, but I'd recommend not going for more. For the scene introducing the new boss you could have: 1) The scene should establish
as a threat. 2) The scene should communicate at least one aspect of 's personality/style. 3) The scene should communicate one potential weakness of
TL/DR: The Should is the answer to how you approach getting the must. If Must is introduce a new NPC, Should is that the experience should be scary. Executing well on Should makes a scene better, but never determines success/failure of a scene because you have no control over how a should is received by the players.
Could is less about the scene in play and more about ways the scene could go. This is giving yourself the first few steps of planning out for different ways you expect the scene to play out and ways it could twist.
If a scene must introduce a new villain, and should establish them as a threat, then you should have at least two coulds prepared. One for what happens when it goes well for the villain (things go as planned for antagonist, what do that look like?) A second for what happens when it goes poorly for the villain (the villain can't get anything done right thanks to dice/player interference, what does that look like?)
You also want to consider some unexpected ways the scene could go, or spinoff scenes that could happen. For example, if the thief escapes the PCs grasp it could turn into a chase scene, a combat, or something else and you should know that.
Again, keep yourself to 1-3 ideas for ways the scene could go. The idea isn't to have all the myriad ways a scene can go once presented to the player planned, but rather to have a few ideas for how things could go loosely sketched out in your head. Give your brain a headstart.
TL/DR: The Could is the answer to a few ways you see the scene possibly planning out based on whether or not the 'Should' is successful, or from other factors you expect to be at play.