Depending on which source you believe, game designer Sid Meier once defined a game as a series of meaningful choices, or a series of interesting decisions. I was thinking of that after a game of Dungeons & Dragons that forced me to refine my thinking about decision points in dungeon rooms. To explain, let me use an example in different variations: Imagine a basic setup in which there is a tiger and a treasure in a cage (let's say the cage bars are magical and don't let arrows and spells through), and there is a lever that opens the cage to release the tiger and let the players access the treasure. Now imagine different variations of this basic setup:

Variant A: The lever, cage, tiger and treasure are all in the same room. Although there isn't a visible connection between the lever and the cage door, the function of the lever is easy enough to understand. The players need to make a relatively straight-forward decision about whether they want to fight the tiger for the treasure or not, or they could try to come up with a different solution, like first organizing a pile of meat before opening the cage. I think this variant offers a meaningful choice and should be fun, but it might be a bit on the too easy, too obvious side.

Variant B: The players come across the lever in one room, but the cage is in the next room that the players haven't seen yet. I'd consider this actually to be an improvement over variant A: The setup rewards players that are careful, because once they open the door to the next room, we are basically back to the more obvious choice of variant A. Impatient players, who just pull the lever without having seen the next room, will hear some noise in the next room, and on opening the door will understand what the lever did when the tiger attacks them directly. The choice is still meaningful, the decision still interesting, and maybe even more so because it requires a bit more forethought.

Variant C: The players come across the lever in one room, but the room with the tiger is at the other end of the dungeon. If the players pulled the lever, much later in the game they will come across upon either an empty cage with a treasure in it, a roaming tiger, or both. But they will not know that this was caused by them pulling the lever. The game of D&D that I played yesterday had a lot of encounters of this variant: The group came across a lot of different stuff to interact with, but in most cases the consequences of that interaction were unclear both before making the choice and after making the choice. There was no explanation of what the things we encountered meant, and no way to find out the history or meaning of it all. I assume that some of the interactions we did caused some of the things that happened to us, but I can't be sure.

Variant D: There is no cage and no lever. There is just a room with a door which leads to a tiger guarding a treasure. This is "old school D&D". And while the lack of a decision point makes this a far less interesting design, I do somehow favor variant D over variant C. Because for me variant C is a decision point with no meaningful choice, which makes the decision not very interesting. In short, I believe that a good dungeon for D&D needs to explain the consequences of their decisions to players. If they don't understand what they did, it all becomes just random interaction with no meaning.