The Uncanny Valley

The Uncanny Valley - It is basically a theory that says humans will accept human likeness to a certain level (about 70%) but if the artificial human becomes too human-like without being human (about 80-90%) then we will be uncomfortable with that artificial being.

Why am I referring to this now? Look at the chart to the left. I would postulate that over the past 10 years, video game graphics have went from the 10% likeness scale to the 50% likeness scale. Now 2004-present, we have been wavering in the uneven ground of 50%-90%, in video games and movies.

I would say that when Final Fantasy: Spirits Within was released, that movie unfortunately fell into the Uncanny Valley of CGI. But now video games have finally caught up. Take a look at a build from the upcoming game Crysis (2007).

As far as I know, these are rendered human models that will be placed in the game in a real time environment. Now where do these models lay on the graph? Just to the left of the valley? Just to the right?

That is, assuming the Uncanny Valley is a stable theory, but I do not believe that has yet to be proven.
However, let's look at another screenshot of another game, Half-Life 2: Episode One (2006).





Once again, assuming the Uncanny valley theory is correct, I think Half-Life clearly sets itself upon the left side of the valley. I would also think most other CGI in these days also set themselves up on the left of the valley.
Remember how scary the Mom and Boy looked in Toy Story? (Uncanny Valley?)
Remember how comfortable it is to watch The Incredibles instead? (Left side of Valley?)
Remember how ugly the Final Fantasy characters seemed even though they were so realistic? (Uncanny Valley?)

I'm just interested to see where video game graphics will go from here. With the technology seeming to range from 60% likeness to 90% likeness, it will be interesting to see where companies put their graphics in the future.