There's an idea in fiction that any given scene moves in a particular emotional direction, usually either from low to high (good to bad) or high to low (bad to good). Occasionally the movement may be from high to higher (good to great) or low to lower (bad to worse) but usually it's just good to bad or vice versa. Thus, for example, if a scene starts with a kids birthday party (high) it may end with the delivery of bad news (low).
While this might seem a bit contrived on the surface of it, this is really just a reflection of the fact that something happened in the scene, and thus whatever the situation was at the beginning of the scene has changed by the end. Changes which do not change emotional state aren't that interesting. If we get good news at a birthday party, then the scene can just feel like so much noise.
This is a pretty well established idea. Mckee talks about it a lot, and Snyder came up with some clever notation to use to illustrate it and Robin Laws has a forthcoming book of beat analysis called "Hamlet's Hit Points" (that I'm really looking forward to) which gets into it. So it was with that inmind I started thinking about fight scenes.
The default fight scene, and particularly the default D&D fight scene, has a predictable emotional direction, from adversity (low) to victory (high). Setting aside fair defeats (as contrasted with forced defeats to move aplot along, which don't have an emotional charge so much as an emotional extended middle finger), this is the model for the vast majority of fights. This lead to me asking myself what a fight looks like if you reverse the emotional charge. That is to say, what does a fight look like that ends on a lower emotional note than it starts?
Answering this question has proven surprising fruitful for me, in that it ends up very quickly crossing into the territory of "what makes an interesting fight scene?". See, once you assume victory, then you need to start asking what else is going on in a fight that makes it interesting, and the answer that presents itself is secondary objectives. That is to say, if orcs attack the princess, the players may defeat all the orcs, but still fail (and by extension, end on a bad note) if the princess dies or is hurt.
See, the classic way to achieve a victory that hurt is to make the fight very expensive. Players get to the far end and realize how many resources (potions, spells, hit points and so on) they used up and the feel much worse for wear. When it happens, this can be a pretty satisfying thing, but its hard to plan for. Vagaries of dice or a given fight tend to make this overwhelmingly unpredictable, and the line between resource depletion and emotionally resonant resource depletion is very hard to pin down. External elements (like the princess) are much easier to manage.
Easier, but still not trivial. If the princess is just an interchangeable part (which is to say, the scenario would be the same if they were protecting the duchess or a box or rocks) then the lack of personal investment tends to mean that it lacks emotional punch, and the satisfaction of winning (killing all the arcs) overwhelms the sense of losing (that the princess was hurt/killed). For it to really work, the players need to be invested. Maybe they know and lke the princess, maybe they're invested in the reason why the princess needs to go where she's going, maybe (and this is most crude) there's a concrete and substantial XP reward for the princess's safe arrival.
All of which comes back to one key idea: fights work better when there's a reason for them that is more involved and personal than "Because the enemy is there". Yes, that takes more work than filling rooms with monsters, but the payoff? Huge.
1 - Perhaps more precisely, any interesting scene.
2 - Where I say "Scene" here you can potentially also say beat, but the discussion of that distinction is its own thing.
3 - I feel this is a safe assumption. It does ignore the old school idea that you can always run away, adding another outcome, but even so I think its fair. There will be more victories (even if expensive) than not in most fighty games.
4 - This ignores the fact that the princess's death may have consequences for one simple reason: that something will make their lives worse is almost meaningless to adventurers, at least in the abstract. The players WANT fights and conflict, and if they get them because the vengeful king is sendign assassins after them, then that's just as cool as finding them in a dungeon.
5 - If the XP reward for safe delivery is lower than it would be to just get in every fight along the route, players will seek out the fights, or at least not avoid them.