Anyone who knows me can tell you that I basically do nothing that’s not related to English in some way. I love writing, I do as many different types of it as possible (essays included. Yes, it’s weird).
So, it’s not really a surprise that, when I play tabletops with friends, I tend to be the GM. Some call it being a control freak. I often do as well. But my first few tabletop campaigns failed miserably, and I couldn’t really figure out why, until I really got stuck into writing. The thing is, the players in an RPG will very, very rarely act like characters in a novel, and they certainly won’t act the way you planned the characters to act. That’s why I have a huge amount of respect for people who are genuinely good at GMing – not just competent, but good enough that, 90% of the time, their campaign will be a success.
I can not and possibly never will be able to say that I’m one of those people, but I’ve learned a couple of things since my first campaigns that make it a lot easier.
So, your characters went to the Wise Old Man for advice, but instead of taking your carefully cryptic remark to mean they need to find the Bandit King, they’ve decided the area is haunted by the ghost of a sinner’s crow.
At this point, you kind of have three choices. You can roll with it, and actually make there a spectral corvid to be found. Maybe it’s a sidequest, maybe it’s just a link to more information, so it doesn’t really matter who dispenses it. This approach speeds up the plot, and results in less faffing around, but it won’t work if the Bandit King is actually a huge plot point.
You could allow them to search for the crow until they realise it was a mistake, and they might then guess that there’s a Bandit King. This will slow the adventure down, but if the players are having fun, they won’t regard it as wasted time.
What is a bad idea is just have them find more information that leads to the Bandit King that has no obvious attachment to what the Wise Old Man said. And certainly don’t but in and tell them they’re wasting their time. The former will leave them thinking that there’s a plot point you’re not letting them get, something that you’re not letting them do. It feels like a loose end, and it’s profoundly unsatisfying. And the latter … do I really have to say why that’s a bad idea? Really? It’s just … selfish. They’re not your characters, you don’t control them.
Whatever choice you take, you’ll eventually come to a point where your characters just do something you weren’t planning for. It may even take the adventure in a totally different direction to what you originally thought (for example, your players realise too early where the town with the villain is, and so skip over a town in between, never meet an important NPC and therefore will run into trouble later), and that’s when the creative rearrangement comes in. Shift the NPC to somewhere else, give them another way out of the trouble, something, just so long as they’re not completely stumped by rearranging your plot. If the plot is planned all the way to the end, you’ll end up losing a fair chunk of work. So, if you can manage it, don’t plan everything out in advance. It will only make it harder to let go.
I wanted to think of some other things to say that were also smart, but then I realised that that’s probably the most important thing for me to say. Recognise the players will not do what you expect, and learn not to mind that. You don’t need to create exclusively sandbox worlds, just let them retain their sense of agency.
Oh, and while you’re at it, one of the best ways to create a difficult puzzle is to not know the answer yourself. Less stress for you as well – just wait until the PCs come up with an idea that sounds plausible/too cool not to reward, and go with that.