Puzzle design is one of the parts of The Last Cartographer that I keep revisiting because it is very easy to make something that is technically a puzzle without making it satisfying.
A good puzzle should make the player say “of course” after the solution, not “how was I supposed to know that?”
The opening cemetery gate is a good example. I did not want the first obstacle to be a glowing button that says OPEN GATE. The player needs to look around, find something useful, experiment, and understand that the game expects them to interact with the environment. That teaches the language of the game before the larger mysteries begin.
Thomas’s study goes further. The desk lock, dates, notes, maps and physical details are meant to make the room itself part of the puzzle. Information should matter because the player noticed it earlier, not because a pop-up suddenly announces a four-digit code.
The underwater material has been even more interesting to design. I have been working on sequences where mistakes have consequences without turning every wrong answer into an instant death screen. Air can become part of the pressure. A wrong dial can vent precious time. An injury can create a second problem that has to be solved under stress. The important part is giving the player warning, feedback and a chance to understand what went wrong.
I also deliberately add wrong solutions. Adventure games are more believable when the world reacts to a reasonable bad idea instead of ignoring it. Sometimes the result is a clue. Sometimes a companion points out why it will not work. Sometimes it is just funny. Occasionally it is extremely unwise.
The standard I am aiming for is fairness. The clue should exist. The player should be able to understand the relationship between clue and solution. The answer can still require thought, but it should not depend on reading my mind.
When a puzzle works, it becomes part of the story rather than a gate placed in front of it. That is the goal.