Induced Human Behaviors as a Litmus Test for Simulations

Your big toe:  formerly a thumb, essential for balance, and second-favorite target of competitive epee fencers across the known world.  Big toes are critical to epee fencing.  See, unlike foil and saber, epee fencing closely resembles actual dueling, in which any part of you can be bloodied.  As a result, epee fencers tend to favor a forward-leaning, wide-legged half-lunge, as if to hide the entire body behind a three-inch bell guard, and will strike at anything that wanders into range with the decisive ruthlessness of a scorpion.  What parts of you are closest to that stinger?  The weapon arm is at a height with it, so it is closest.  Knees, while also a frequent target, can be leaned sideways or backwards, as appropriate.  But your big toe is where the rubber sole hits the road.  It can't move much unless and until the whole body shifts.

We can model the physics with the same formulae, measure the anatomically possible response times of muscled joints, and faithfully re-create the behavioral responses to stressors.  But that doesn't guarantee our simulation will be "like the real thing".  The reason is input:  not even a steering wheel controller accessory has the leverage, inertia, or feedback of a real steering wheel, so when mapping a handful of buttons and a couple of sticks to a human body, we must compromise and sacrifice.  But the diminished system still faithfully simulates if the same human behaviors emerge.

In a fencing simulation, do real-world fencing tactics succeed -- and fail -- as in the real thing?  Does the game create an inclination in the player to feint high, then deceive to a low line?  Can a competent player regularly riposte an amateur's machine-gun stabbing?  Does leaning back to outdistance a lunge by a mere inch give the defender many good options for the counterattack?

It's an oft-overlooked fact in racing games that moving the analog stick to the maximum position does not, in fact, turn your car's front wheels to their maximum.  Rather, it turns the wheels to the maximum they can turn without the car skidding or flipping.  This polite interpretation allows the player to concentrate on the race itself, while a direct stick-to-steering connection would not.  That same feature didn't exist in Rally Cross, and damned if it didn't faithfully recreate off-road races as driven by a bunch of over-enthused hillbilly amateurs, tumbling around corners and up walls.  It was the most fun to be had while losing.

When a player learns via simulation that a powerslide is the result of having too much weight on the outside back wheel, so the traction cannot hold the car and it slides, plus having not enough weight on the inside back wheel, so the traction cannot hold the car and it too slides, then the game has given him some take-away value that an easier game with a Powerslide button could not.  When a player learns how to finesse longer-limbed opponents with ripostes and deceptions, it proves that matching the induced human behaviors, more than the actual physics, ultimately determines whether a game is like the real thing.   So stay on your toes:  because after wielding a Wiimote for the fiftieth time, then an epee for the first, she's still playing the same game.  Only with a more intuitive interface.