Dynamic adventures are a bit different from encounter based ones. This is not to say they don't include encounters - they do - but the priority is less about a specific set of encounters and more on the encounters that will emerge organically from the events of play.
The primary thing that defines a dynamic adventure is that the characters are usually proactive in pursuit of a larger goal. In an encounter game, players may be proactive in the short term (that is, they may proactive seek to get to the bottom of a dungeon) but in the broader sense their actions are directed by external forces (like employers) or simple motives (profit).
This line can be a little fuzzy. An encounter based game may have character with personal motives providing subplots, but in a dynamic game, the entire direction of play is shaped by what's important to the players. To illustrate, consider the case of the Lord of the Rings - it would be entirely possible to play it either way. As an encounter based game, the players would have a goal (DROP RING IN VOLCANO) and the GM would seed the map with encounters that the characters would encounter as they go. In a dynamic game, players would have a subtly different goal (DEFEAT SAURON) and they will find a way to accomplish this. Maybe they drop the ring in the volcano, but maybe they do something the council rejected, like take up the ring themselves or sail across the sea with it.
Some GMs are nodding here, and others are wincing, because this is an invitation to completely break a campaign. In an encounter based game, this is a problem. In a dynamic game, it's an *expectation*.
Still, it's not just a matter of waving a wand and saying to your players "Go nuts, dudes!" You can do that, certainly, and that's very nearly the definition of empowered sandbox play, but that's another topic entirely. For dynamic play to work, the characters need a clear but open ended goal to drive play. That seems like a simple enough thing, and in cases like DEFEAT SAURON it pretty much is, but it can be rough to come up with something that works.
Of the two criteria for a good goal, clarity is probably the more straightforward. It needs to have a little specificity to it - notice that DEFEAT SAURON is a specific verb and noun combo, while DEFEAT EVIL would be pretty vague unless there's actually a guy named EVIL in the campaign. In a pinch, the VERB NOUN model works pretty well, though. I mean, you can make it longer with fancier language, like EXACT REVENGE ON SIX FINGERED MAN WHO KILLED MY FATHER, but practically that's PUNISH 6-FINGERED MAN. Specific, actionable verb, and a specific target. Perfect.
Notice the frequency of the use of the word "specific" there? That's the trick of it. A goal is clear when it is specific, both in its outcome and in its call to action.
The second criteria, that it be an open ended goal, takes a little bit more thought. The goal should not be immediately achievable and it should not have an obvious course of action. "Rescue the Princess from the cultists" is a good goal for an adventure (because, hey, find the cultists, attack the cultists, rescue the princess, done!) but it's not going to invite much variety of play or drive a campaign.
In contrast, if the players had a goal of DEFEAT THE CULT, and the cult captures a princess, then the short term goal grows naturally from the big goal. One of the signs that you have a good goal is that it's easy to imagine secondary goals that spin off it for the sake of a session or even a few scenes.
Once you have a goal, play has a direction. Players with a clear, open-eneded goal are usually readily to launch out of the gate, so you need to be ready for them, and that's what the next part of this will address.
1 - Sandbox Play is a term for a playstyle where you can go anywhere and do anything you can get away with. The idea is that there's no game per se, just something to explore, akin to the difference between a toy or a game. In some sandbox games the players are observers, changing the world very little. In others, the world is utterly transformed by their play. This style is fantastic for exploration games, but it can be a little unfocused (for better or for worse).
2 - Maybe more than one. How many goals a game can tolerate depends a lot on the specifics of the group.
3 - There's also a third secret criterea - that it be a goal which engages the players. It is often quite reasonable to start a game without a goal and to instead find it in play. A goal found in play, especially if it involved someone really angering the players, can get you more excitement and engagement than the most planning.
4 - If there is, that may suggest you're actually playing a Japanese video game.
5 - Unless you are specifically limiting the scope of the game. In that case a goal which can be resolved in one session is totally in bounds.
6 - I still haven't defined KWORC yet, have I? Man, I'm a stinker.