All very fine examples, but at the same time they give their own reason why people watch it: They wait for something interesting to happen. Some twist, something exciting or simply the ability to say "gee, at least I am not that deep down in the shit".
That's also the appeal of talk shows, I'm kinda convinced.
I don't get Dirty Jobs, I guess it's one of those "I'm glad I don't have to do it" things. OC Choppers I chalk off as a mix of the hope that the old guy freaks out and of course the appeal of stuff that has some sort of engine in it to some guys. Big Brother is simply... heck I dunno. Maybe a mix of all those traits.
Nothing of this applies to games. It's a given that there will be fights, but they are uninteresting because nobody can really die (which is, btw, IMO the appeal of those "Our soldiers in Afghanistan" shows that pop up left and right right now). Plus, you've seen one fight in most games, you've seen them all due to the limits of the game mechanic. There are only so many "moves", only so many animations your character can go through 'til you have seen them all. The only appeal afterwards is just that you, the player, have to get better at executing them, something that doesn't apply to someone just watching. The story of most games is mostly different from the average "Sally does Houston" in terms of the ESRB Rating, but it sure matches the amount of surprising twists and the thickness of the plot. Which may be ok if you're part of the plot, but I couldn't think of a game the story of which would be interesting as a mere spectator. Plus, considering the speed it progresses usually is so slow that you'd wish someone dug out the Dune director's cut and added some padding so you had something to watch instead.
Let's be level here, most games are simply not interesting for spectators. They were not meant to be. They were meant to be interesting to someone playing them. The problem here is that if they were interesting to watch they most likely wouldn't be too interesting to play. Because that would require an awful lot of putting it on rails, for more than one reason.
First and foremost, because then most people could not figure out where the heck the next step should take them. Ponder any movie that may come to your mind and tell me that you would have foreseen what the hero does next. There are simply so many information gaps that would have to be filled for the player so he could possibly come up with the same idea. Well, either that or he'd have to have played it through already so he doesn't do heaps of wrong turns first (which are, as we can hopefully all agree, not too interesting to watch, who'd want to watch a guy hit the "you cannot progress here now" wall for a few hours?), which limits the whole thing to walkthroughs, and we already got plenty of those.
Then there is the limit of paths the programmer can offer you. If the game offers you nearly infinite ways to progress, by the laws of the market they have to be very short ones. They can only put in so many "story lines" in a game before they run out of time (or money), which means that instead of, say, 30 steps to success they offer you 3x10 steps, which invariably leads to a shorter game experience.
And that's the main difference. There is very rarely a "surprising twist" in a game (and if, a 5 minute video showing just that is plenty more exciting than watching someone hack through 10 hours of monsters to get there). Most games are incredibly repetitive, which can be interesting for a gamer if the way it is presented and how he has to react to it is interesting, but it is mind numbing boring to watch. Take Elder Scrolls. Yes, playing it is fun. Making someone watch play it should be chalked off by AI as torture.