This is a for instance, using baseball. You are the centerfielder. When a ball is hit, a trained human can predict with great accurracy where the ball is going to hit, based on many things a computer simply can't quantify. Example, who is hitting the ball, where are they (the hitter) in their career, the game situation, maybe he just needs a base hit, the wind, the sound of the bat hitting the ball, is the grass wet/dry, the direction the player was facing when he hit the ball. A computer needs hard facts to make it's cecisions, humans can fill in the blanks based on situational data and experience. Minda you, while your brain is doing all those calclations on the fly, it is also monitoring millions of other calculations that keep you alive. There is your Super computer.
Your view is in fact the dominating view among sicentists: the brain as a super computer. Still, even if they are right, there are computer configurations which -theoretically at least- instead of using a digital model, use so-called neural networks that emulate (parts of) the brain. There is at the moment, not a single computer, super or otherwise, that can compete with the human brain. So the example you gave is a very good one.
My view on the matter is a little unorthodox, even if closer to the neural networks idea: I do not think that the brain is capable of any sort of computation. I firmly believe in the following:
You cannot explain intelligence with intelligent parts.
That means that, as far as i am concerned, neurons are just as dumb as transistors. And since brains are all made of neurons, you wonder where the computations can take place. Hey, you will say, a computer is all made of transistors, and still, it is able to compute. Okay, let's look at it more closely:
Computer instructions are a series of switches that have to be turned on/off in a specific order. Have you ever watched the Japanese domino shows, with the beautiful figures? That is exactly what a computer instruction is like. You hit the first switch, and the rest follows as if it had its own will. A computer is a super Japanese domino show, with millions/billions of domino configurations, and not a single domino can be said to be smart. Every time you load a program, you are placing domino's in a certain configuration, ready to stumble at your command. The only smart thing in a computer are the Intel/Amd boys and girls which have prepared all those configurations for you to play with.
Researches who believe more in neural networks also make the mistake of thinking in terms of computations because they think that is the way brains work.
Allow me to simplify your example and take the old computer game of ping pong (or whatever it was called). A very simple game, the ball appears at a point up in the screen, and you have to intercept it with your bat/cursor, by moving it to the right position.
Is the brain computing the ball trajectory? I personally do not think so. It takes years for a human child to master his hand and eye coordination, and professional players have to train everyday if they want to keep their edge. Why is that?
Maybe because the brain does not have any magic formula that tells him where to move the bat, and the only way he can get better at the game is memorizing each position and the corresponding actions.
