I feel it is a great mistake to believe that it’s there to challenge the characters. They’re imaginary. They’re not playing the game. It’s the players sitting around the table with you who the puzzle is for. And, if your players don’t like puzzles, it’ll be a horrible encounter.
Imagine describing a logic puzzle scribed on a door. To solve it, at least one character must succeed on a DC 16 Intelligence check. Is that a puzzle? No, it’s an ability check. Is it interesting? Not very.
It is entirely possible to view D&D combat as one big puzzle. The reason? You’re making decisions and thinking about it all the time. And that’s at the heart of any good game: You need to be challenged. You need to make decisions. Games falter when there aren’t enough decisions or they’re boring or pointless ones. I’m looking at you, Candy Land and Monopoly!
Are you interacting with the King? What do you say to him? How do you persuade him that his daughter wants to run away with you and become a dragon hunter?
Different players have different levels of abstraction with such matters. You have people who want to act the entire scene out, and you have people who just want to let a single ability check determine the result. And you have people in-between. None of these approaches is wrong, but they prove very difficult for adventure designers. You’ve got to cover a lot of bases. However, because the entire situation depends on the players making decisions, it is interesting. The key decision? “We’re going to persuade the King his daughter should come with us, rather than just leave her behind.” That’s what makes it interesting. You have a decision, and it has a consequence. The delight about a RPG is that you aren’t sure of what the consequence will be. “Well, I’m actually a dragon,” says the King!
There was a mechanic in the 4th Edition of D&D called the Skill Challenge. The idea was that to succeed, you needed to make X successful skill checks before you failed 3 times. The choice of skill was mostly determined by the challenge itself. And, for the most part, they were deadly dull. The problem was that the mechanics by itself wasn’t interesting. It boiled down to the players choosing their best skills and making rolls. You could make a little puzzle out of it which was “which skill do you use?”, but even that was mostly unsatisfying. They were just too simple.
Consider a D&D combat. That is full of interesting decisions. Where do I move? What do I attack? What weapon do I use? Do I protect the wizard? Do I cast an attack spell or heal the fighter? Can I remember the secret weakness of this creature?
You don’t want every challenge in a D&D game to be that complicated, but you do need challenges to be complicated enough to engage the players.
An individual ability check works when it’s made in context and as part of a greater tapestry of decision-making. Requiring a Strength check to open a door isn’t interesting by itself, but when it might require the players to enter the dungeon by another method, or alert guards to their presence if they fail: that makes it interesting. The “logic puzzle” that is solved by an ability check? It’s interesting if it has context. If all the players chose to play low-intelligence characters? That makes it interesting – it provides a consequence for their decisions during character creation. Perhaps they can find a sage to give them the answer?
It’s when there’s only one solution to a challenge important to the story and the players fail at that challenge that you get real frustration. I like some puzzle challenges – the ones I can solve easily. If I’m sitting 30 minutes later with no solution and the DM is enjoying the fact we can’t reach the end of the adventure? That’s not fun. Likewise, I have a lot of problem role-playing when I don’t know enough about the world I’m in. I prefer to abstract those sorts of challenges out a bit – describe what I want to achieve, and allow the DM (perhaps aided with dice) to determine how I do.
I’m happy to have a secret door that the characters might not find if it just leads to a bonus encounter. If it leads to something that the adventure falls apart without? Not so good.
Ultimately, the DM should be guided by the desires of the players at his or her table. Don’t throw combat after combat at them if they don’t enjoy constant combat. Don’t set them a high-level logic puzzle when they don’t want to do one. They enjoy role-playing and making ethical decisions? Well, perhaps D&D isn’t the best game for them, but I’ll give it a shot – it’s a very flexible game!