Via MeFi comes an incredible 11-minute complete playthrough of Super Mario Bros. 3 (18MB Windows Media file).
I was wondering why watching this movie's so satisfying. I've reached the conclusion that it's rather like knocking back your first double espresso after a lifetime of sipping Coca-Cola.
Here's an interesting question: why is playing a Super Mario Bros. 3 so much more fun than, say, balancing your checkbook in a spreadsheet, which also involves manipulating little electronic symbols on a screen with a sequence of button presses? It's because SMB3, like all great games (video or otherwise), delivers a constant stream of little sensory rewards to the user for completing certain actions --- the jingle that plays when you earn a 1UP, for example, or the "plick!" sound (accompanied by the visual kick of seeing Mario double-jump) when you step on a Koopa, or the resounding "ding-ding-ding-ding" of grabbing a line of coins. The simple aesthetic pleasure that you experience when you receive these rewards make the game rather like a drug: each little victory is like a "hit" of the drug that keeps you playing in order to score another one. Which is why the ultimate compliment one can bestow on a great game is to call it "addictive".
When you string together a bunch of these victories --- making Mario leap through a sea of coins, bounce off a string of enemies, and then fly into the sky on his raccoon tail --- the "hit" becomes a sustained high, the dime-store cousin of watching a virtuoso performance in music or professional sports. Watching someone play through an entire game flawlessly is like mainlining a clean pure gram of the drug that you've only tasted previously in watered-down and adulterated form. (The experience is only slightly diminished in this case by the fact that this movie was produced by playing back a recorded series of button-presses on a SNES emulator.)
As I've said before, I've basically quit playing video and computer games; but this movie brought back fond memories.
UPDATE: As I was writing the above, I had a dim memory percolating around my brain about something I'd read that discussed reward structures in games, but I couldn't in it down. I've finally tracked down the reference: see Daniel Cook's thoughts on "reward schedules" and John Hopson's article on evolutionary game design.