The idea may not be, but the moment it becomes a demo it can be worth quite a lot. The story, characters and gameplay can be very valuable, it is how the idea is sold to publishers. This is just like scripts in the movie industry. Most publishers have internal studios, that could possibly be sitting idle. While it is not often done, there is the opportunity for a publisher to move on a developer's idea much faster than the developer could.
The video game industry is brutal, especially with people doing copy cat projects. Look at the current farm games on facebook. The first one was FarmTown, made by a small group. Now some of the facebook powerhouses have released the exact same game with nearly identical gameplay and even similar art (FarmVille from Zynga and Country Story from Playfish).
Or just look at the rivalry between Rock Band and Guitar Hero.
The execution is still worth much more, but I think there is still a need to protect the idea.
Also, it's not like FarmTown is getting destroyed in the marketplace after it's been cloned - they're still the third most popular game on Facebook with 18 million monthly active users (behind FarmVille and Mafia Wars).
Sure, they probably were not the first ones to come up with the idea of bringing a social farm game to facebook either. However, it is clear that once Zynga and Playfish found this idea they could rapidly turn it around too.
It is a bit scary as a small developer to know that you need to share your idea with publishers, but to also realize that they can turn around and dump so much money on your idea that they could produce something much much faster that you cant even hope to compete with them.
It is just a different climate than web based startups. Console game especially are pretty much not possible without publishers, which is not something a 3 man web startup ever really has to worry about.
If VCs were keeping tech teams in house that could potentially jump an idea i think the startup community might see the value of ideas a bit differently.
The Rock Band / Guitar Hero situation doesn't fit your description. Harmonix developed Guitar Hero. Red Octane makes the guitar peripheral and owns the rights. Activision bought Red Octane, and along with it, the Guitar Hero brand. MTV Games bought Harmonix, and with it the developers who then went on to make Rock Band.
This is a situation where those responsible were split in two, not one company looking over the shoulder of another.
Coincidentally, the Ludum Dare contest just finished (I didn't enter this time.) Here's a crapton of games, all written in less than 48 hours, and all themed around caverns! :
I thought this would be more interesting, but it's just a rehash of the basic "execution is key" message; the "game ideas" aren't even game ideas so much as generic genre themes. They could just as easily work as pitches for books, movies, or tv series. The mechanics are the message.
I've been cranking out reasonably full-size games every 8 weeks with a few guys for Blurst.com, so we're in the idea generation phase pretty frequently (Off-Road Velociraptor Safari, Minotaur China Shop, Time Donkey).
As a general statement I think it's fair to say that game ideas are worthless. If you want to get semantic on it, though, I think the problem is that most people think theme/treatment/context ideas are game ideas ("You wake up in an insane asylum with no memories").
If you define a "game idea" as something that clearly states what the player is actually doing, from second to second, then I think game ideas have worth. People who demand NDAs just to hear their brilliance as still full of shit, obviously, but ideas distilled to the brink of execution have merit, and can be bandied about in discussion in a useful way.
Of course, boiling a game idea down to that level of granularity is quite hard, so if you have that done you've done much of the mental clarification already.
At that point you may be talking about a game design, which is more than the idea. Just as there's a long way from the idea of a program to the program itself, there's a long way from a game idea to a game design.
The mistake people make in framing this conversation is that while some people come up with high concept ideas and don't think about the execution. Most people at least, myself and other professionals I've worked with(In game development or web development) tend to come up with clever implementation hacks as well as ideas. And really it's the cleverness that you don't want to share, often you've spent much time figuring out and collecting economically creative implementation details. It's these that we worry about sharing easily, and really are what make an idea feasible beyond it's artistic merit.
I don't favor keeping all things to myself, just wanted to frame the debate more succinctly to discuss the aspects of the debate separately.
Although I agree with the article's overall idea, his example of Okami and Legend of Zelda as games with totally different ideas and similar game mechanics is a horrible example. Okami's game mechanics are quite different from Zelda's.
There must be some game ideas which are worth plenty on their own. Tetris, for instance. Or... well, actually that's the only example I can think of. But it's a big one.
The video game industry is brutal, especially with people doing copy cat projects. Look at the current farm games on facebook. The first one was FarmTown, made by a small group. Now some of the facebook powerhouses have released the exact same game with nearly identical gameplay and even similar art (FarmVille from Zynga and Country Story from Playfish).
Or just look at the rivalry between Rock Band and Guitar Hero.
The execution is still worth much more, but I think there is still a need to protect the idea.