Interaction leads to immersion when the player's actions maintain the illusion of reality that is largely at the core of a fully immersive experience.
Someone used Wii Boxing as an example, but as someone with real life boxing experience, I have to say that the control scheme for that particular game was extremely frustrating as the program simply does not account for speed and reach (or even basic footwork); two of the most important factors in allowing physical ability (or physical build) to produce a match winning advantage. My actual punching speed meant nothing in Wii Boxing, which became obvious after the first few minutes of play. That was a disappointment since I was really looking forward to a shadow boxing game with a virtual sparring partner. It did not meet my expectations.
As far as creating the illusion of boxing, Wii Boxing failed miserably. Eventually I discovered that "shaking the maracas" was the most effective way to throw punches in that game and hit "pro" status.
If other factors in a game are able to create an illusion realistic enough to make the actions of your on screen character "your" actions, then life like visuals are not essential.
Remote slashes standing in for sword swings are far more satisfying that button presses. NMH added a bit more to the repeated controller shake for a flurry of swings (like Zelda TP) by adding specific directions for finishing blows. Unfortunately, the big arrows on the screen take away from the illusion, but there are limitations to what can currently be done.
Pointing the remote as opposed to using the right analog stick feels more natural and satisfying. Aside from the fact that you're shooting with what feels like a TV remote, it does add an extra degree of immersion.
But realistic visuals are particularly vital in maintaining the illusion when it comes to objects, environments and physical actions we are very familiar with. Fantasy worlds or science fiction worlds are not held to the same standards as say a walk through NYC (GTAIV) since they generally don't exist in the real world, subject to real world physics and visuals. The only thing we have to compare in the sci-fi/fantasy worlds are what we see in film and television.
But driving simulators for example, will feel a lot more immersive the closer the visuals are to the real thing. Having a convincing cockpit 1st person view along with realistic engine and tire screeching sounds accompanied by controller feedback through the rough sections, etc. all help to maintain the illusion of driving a car.
Ultimately, it's any combination of factors that succeed in pulling the gamer into the world created by the developer. As long as that spell of illusion is maintained, any game can succeed in creating an immersive experience, even if the interaction is limited by button presses, key stroke presses, or arm waving motions. I'll go so far as to say that the biggest factor in creating an immersive gaming experience is a solid story line with a good script, well developed characters, set in a believable world, even if that world is a sci-fi or fantasy world.
As long as the player continues to feel his or her actions are shaping the events unfolding throughout the game, that player should be immersed in the game.
But until we start seeing one for one action translation on screen like a holodeck, where your body and any objects you have on your person are the "controller", control schemes alone are not enough to create the same level of immersion.















