Highlighting paths
An important difference: to indicate the path is one issue, the other is to know the path to the objective. For example: if you are going to climb the Everest there are multiple routes to the summit, some harder than others, but they all lead to the same objective. From many Himalaya's spots it's possible to see the Everest's summit, the issue is how to get there. Many games do that, they show the destination far away on the horizon, but the player has a long journey ahead with many obstacles to reach it. To identify the path and to identify the destination are two distinct tasks and one doesn't depend on the other.
Landmark: a landmark has two purposes. One is architecture, give the environment a theme to be built around. The other is to give a direction for the player, who can use the landmark to guide themselves. In a city what are the landmarks? The Christ the redeemer in Rio de Janeiro, the Central Park in New York, the Eiffel Tower in Paris. Large buildings, parks, statues, bridges, are commonly landmarks. Natural landmarks include valleys, mountains, rocky formations, lakes. If you are in a desert region, in the middle of the ocean or in a dense rainforest there aren't any landmarks that you can follow and that's te reason for people getting lost in such places. In a game it's undesirable to get lost.
In a game the landmark doesn't necessarily has to be a large building. It can be a tree, a painted wall or a flag. It's not always a matter of size.
Path or route: in closed spaces with many corridors and turns to the right and left the player becomes disoriented because there is no landmark to guide their selves. If there are windows that allow the player to see some fixed point or area it becomes much easier to avoid walking in circles. If the architecture doesn't allow for windows the solution is to have a map or to place lights, lines or a character to follow.
There isn't a perfect way of indicating a path, but it must be easily recognizable in the environment. The most important concept is that the player must be able to see the path. Can there be a unexpected route? Yes, but it can't allow the player to skip parts of the game as this would mean that the level design has an error. To facilitate the player's navigation in a level avoids frustration and headaches.
A huge flaw in bad level design is to not properly guide the players, leaving them lost. Worse still is a reversed design that highlights paths where there is nothing of importance and the path that the player should take isn't properly highlighted. In a game with a certain objective to be completed the lack of guidance leaves the player frustrated trying to find its way. In a multiplayer game the players have a hard time trying to find each other.
In some games there are cheat codes to enable god mode. I've found something quite interesting with it. If you play in god mode and find yourself lost in a level, this means that the enemies were placed strategically to guide the player. How so? If you are invulnerable to all damage, you tend to disregard covers and not care about where the enemies are comming or attacking from. The same idea could be applied environmental hazards. If the player can run straight ahead without caring about taking damage, this could mean running away from the intended path.
Footnote: depending on the game's context the lack of a landmark or a clear path is intentional and there may be a good reason for that. Finding the exit may be the challenge itself for instance. In the case of secrets or shortcuts, to hide the path is desired.
Nightclub Kiss, Santa Maria, 2013: a fire in the nightclub killed hundreds of people. Among the deaths many were found in the club's bathroom. Why there? When the fire began the whole place was filled with a thick black smoke that prevented everyone inside from seeing the escaping routes. Amid the ensuing chaos the bathroom's bright green light caught the attention of many who were trying to escape.
Challenges and obstacles
Difficulty: every game has to have a certain level of challenge, it's the most basic concept of all. But there is a threshold between challenging the player and lacking good sense. The obstacles should be placed in an ascending order of difficulty, beginning with easy obstacles and ending with harder obstacles. If the obstacles are placed in a random order the player is confused and the game's progression is harmed. It feels anti-climatic to have harder obstacles placed before easier ones. Obstacles that rely on pure luck and are random in nature shouldn't be used. A player wants to feel rewarded by his own strengths, not punished with randomness.
The challenges can't / shouldn't be all of the same type. There is room to be creative and the same ability can have different purposes in different contexts or the same type of challenge could be solved with different abilities.
Traps: A good trap makes the game more intelligent, more interactive and/or harder. If it makes a game excessively hard or its only purpose is to hinder the player, then it has lost its purpose. Any player identifies a pattern over time after coming across traps multiple times. If a trap is randomly placed or doesn't follow a pattern in the level design, it leads to confusion and frustration and turn into hindrances. The worst thing a game can do is to place traps that intentionally spoil the player. Traps that are meant to make the player err. This annoys players and shouldn't be done.
Intuitive obstacles: some obstacles don't require explanation because they are self-explanatory. That's the case of fire for example. If there a pipe leaking with fire on it blocking your way, every player knows that fire burns. If the player has to go through there it's natural for the player to look for a way to stop the leaking to put out the fire. If there is wind pushing the player and a certain wind tunnel can be used to thrust the player over a long jump, this should be intuitive for most players as wind is an universal thing.
The problems arise when there are obstacles that are not intuitive. For example blowing up or destroying a door. If the game presents you with the means to destroy that door and it clearly shows that you have to destroy that door, that's intuitive. However, if the game forces you to guess that you have to destroy that door and doesn't give you any hints or clues about how to do it. That's not intuitive. Suppose some part of the game requires a password to progress. No player is going to guess the password or try out all combinations one by one. The game must provide some hint about the password. If the password is placed in front or at the side of a door requiring it, then the obstacle has lost its purpose. Suppose there is a multiple choice question and there is only one chance allowed. The player has to rely on knowledge about the question to choose the answer and that knowledge is somewhere in the game. If the game presents you with multiple choice answers and no question at all, how are you going to guess what the question is? In an extreme case there is neither a question nor multiple choice answers. The player is left with no tools, no directions, no hints or clues, nothing. That's like a doctor attempting to diagnose a disease without knowing any symptoms.
Counter-intuitive obstacles: one of the worst things that a game can do is to be counter-intuitive. There is a universal language related to human perception such as light not going through walls, unless the wall is transparent. Ice represents cold. Gravity pulls everything downwards. A game can have good excuses to not follow those, but the excuses have to be feasible. A game asking to turn on the lights by destroying a power source is counter-intuitive. Another case for that to happen is when a common object is required to do something that doesn't match the purpose of such object. It can be done if the game is resorting to creativity, otherwise it leads to illogical scenarios.
In many ways intuition is closely related to physics. Take gravity for example. Two objects, one being much heavier than the other, fall with the same acceleration and speed. Being heavier doesn't make something fall faster. It's a matter of aerodynamics. To calculate physics in real-time is a computationally heavy task and oversimplification of physics can lead to obstacles or challenges being counter-intuitive. Very often, two boxes, one made of cardboard and another made of wood, feel the same in game because they have the same mass in the eyes of the physics engine. That could be an artist's choice that mismatches the physics in game. Nonetheless, breaking a glass window with a paper ball wouldn't make sense, irregardless of physics.
Contextualization: the obstacles and challenges must be part of the level design. The obstacle has to blend within the environment to not feel out of place. In the Indiana Jones movies the traps are made of ropes, rocks, arrows. There is nothing electrical or electronic devices because Indiana travels to ancient places that lack electricity. There is also a matter of the most basic physics principles and as such, light cannot go through opaque walls. The glass is not strong enough to support the weight of a building. Ice is slippery. Paper catches fire. So on.
If there is a river that is blocking the way and the river wasn't there before, then something happened to create that river. It is natural to think that it has rained, a dam must have broken or some explosion changed the course of a river for example. A door blocked by debris, the debris came from some explosion. It feels detached to have a military base with a door that won't open due to a short circuit when the whole environment is clean, new and with no apparent destruction.
Quality assurance: the developer must see the game from the perspective of a regular player. On the other hand, the player has no obligation to guess what the developer wanted to do. At the same time the developer must not doubt the player's intelligence nor insult him with challenges that have no solution or leave the player with no directions.
A challenge stops being a challenge when there is a unwanted trivial solution or when there is a way to bypass it. The challenge must work as intended. If it relies on a certain sequence of events, depends on time or specific conditions, there should be margins of tolerance for errors. If there are conditions in which the challenge can't be solved they should be predicted, otherwise the game won't end.