Infiltration is a minority genre in video games. Yet more and more games include infiltration components / way of playing the game, like recently Deus Ex: Human revolution, or Batman: Arkham Asylum. But there are praticly no longer full games of infiltration.
We could say that infiltration is divided in 3 important steps:
Step 1: Observation: Knowing the environment, the ennemies, their paths, the hideouts...
Step 2:Plan-etablishing: Now that you know the place, thinking of what to do, who to kill, in what order, etc...
Step 3: Plan execution: Now it is time for action. Acting as you planed, and it may work... or not.
As a big fan of the genre, I've come to play a large amont of infiltration franchises, from the first Tenchu in 1997, to today's Deus Ex: Human revolution. In everyone of those, good or bad game, there are mistakes that still persist. So I decided to list the most commons.
1. The "Easy-fighting" mistake:
The purpose of infiltration game is to avoid front confrontation with ennemies. Satisfaction comes when you pass through a room without being spotted, whether you knocked down NPCs or simply passed through. Full satisfaction is gained when you though of a right plan (Step 2) that perfectly worked (in Step 3). It is a reward for your mind, somehow similar to a puzzle game: you discovered one right answer to solve the problem (current room).
If you get spotted, that means either you haven't found a right solution, or somehow your plan did'nt work so well. So the game punishes you, and if you don't hide quickly and well, you may get killed.
If direct confrontation becomes too easy for the player, the satisfaction will sensibly decrease, because the player will think "What's the point of thinking so much if I can just go and kill". Exept for hardcore-gamers who challenge themselves. And if you are designing a cross-genre game.
2. The "My buddy's dead? Nevermind" mistake:
(This is, I think, the most stupid and frequent mistake)
Because you can't take a hundred bullets in the face, infiltration tends to be realistic. Not in the background or in your abilities, but in the weaknesses of your character. This is an important part of immersion. But this realistic side collapses everytime we see this: When you kill/knockdown a NPC, the body might be found by other NPCs, triggering the alert. So you hide in order to not get killed. And, after a certain amount of time, usually between one and two minutes, the alert stops, some gards return to their HQ, and the others simply continue their patrols.
I mean, "come on guys, you know I'm here and I am going to kill you. One of you is already down... Doesn't it bothers you?". In one minute of game, all of the immersion disapeared. In the Tenchu series, that goes so far that you can imitate a dog barking while a guard is looking for intruders, and the alert will immediatly stop.
I've always wondered if the cat imitation in Tenchu wasn't actually a joke.
I agree the game will become too hard for some players if there was constantly twice as much guards and if they were constantly looking for you and after each other. But still the fact that your plan failed has to have consequences, either the player will have more difficulties, or he will load and restart the room. Otherwise you break the immersion and the satisfaction, same way as in mistake one.
3. The "Obvious one alterate path" mistake:
As said before, a huge part of infiltration is thinking. Thinking of the right path to take, how many NPCs you will knock down, which first... It is a lot similar to the reasonning in a puzzle game such as Portal, just to quote one of the most famous. You have to think about what moves to do, in what order, what it will affect in the rest of the room.
But if puzzle games works with the "one problem-one solution" approach, it is different in infiltration. The problem, "how to pass the room," is constently changing depending on the way you chose to take. It depends of the approach of the player. Is he the type of player that will try to kill every NPC in the room? Or will he simply try to sneak without any confrontation? Infiltration gives a lot of possibilities to player.
Hidden shaft, sneaking desk-to-desk, silently killing everyone... Lot of solution for one problem.
Now this mistake rarely applys to full-infiltration games, but mostly in other games which tries to add a small sequence of infiltration. Here is a meaningful example:
There is a corridor, two guards are talking. If you go ahead you will, no, you shall, get spotted. There is a ventilation shaft on your right, which you could'nt not notice. Good now take it... Crawl inside... Good, you are now three corridors ahead. You avoided confrontation using infiltration. Did you?
You see there the player did'nt though of how to do, or planned anything, he took the only path proposed that would'nt get his/her character killed. There is no Step 1, or Step 2 or 3. It is only a single path, hidden by the impression of another one, who leads to certain death.
These are the main mistakes to avoid while designing an infiltration game. This is saddly one of the fewest represented genre in video games, (as full-genre game). It is also one of those that can evolve the most (especially in IA). Yet today, the genre becomes more and more popular, thanks to cross-genre games that allow differents approaches, strating with the first Assassin's Creed in 2007. Perhaps in few years we will see a boom of the infiltration genre...