game design - 1% idea, 99% execution?

I keep reading in the game development community that the idea of a game is 1% and the execution of the game is 99%.

Do you believe that to be true? Not necessary in those percentages..

I always thought that a great idea is invaluable and it makes a game successful even though the execution is simplistic.

Answer

Why Your Game Idea Sucks

Your game idea sucks. That's the bad news. The good news is everyone's game ideas suck, so you're not alone. And hey, now that you know, you can stop wasting your time worrying about it and actually sit down to make a game...

Also, there's a difference in simplistic and badly done. Pacman was simplistic, and BattleCruiser3K was horribly done (I actually once bought a copy - the cursor movement on the screen, rarely matched the mouse movement). The game's idea was fairly cool, but implemented in a way that was unplayable.

For another excellent example of why execution matters more than the initial idea, consider the history of Quake:

The earliest information released described Quake as focusing on a Thor-like character who wields a giant hammer, and is able to knock away enemies by throwing the hammer (complete with real-time inverse kinematics). At the start, the levels were supposed to be designed in an Aztec style, but the choice was dropped some months into the project. Early screenshots then showed medieval environments and dragons. The plan was for the game to have more RPG-style elements.
[...]
Eventually, the whole id team began to think that the original concept may not have been as wise a choice as they first believed. Thus, the final game was very stripped down from its original intentions, and instead featured gameplay similar to Doom and its sequel, although levels and enemies were closer to medieval RPG style rather than science-fiction.

What made Quake a great game was not the initial ideas, which changed frequently, but that id Software had a great programming team, including people like John Carmack and Tom Hall. They didn't depend on an idea that sounded great - they built it, and tested it. And when it didn't work as intended, made changes. They evolved the game over time, to something that wasn't even close to the original concept.

Update - Just ran across a great article (source DaringFireball) which explains that Ideas are just a multiplier of execution - nicely explained, using dollar values for ideas.

Update 2 - the question seems to have re-surfaced, so here's another data point that comes to mind: The Aristocrats. This is a DVD of 100 comedians telling jokes - and they're all telling the same joke (a very old, very dirty joke called the Aristocrats). But of course, they're all telling it differently - the joke is the same, but their execution is totally different.

Of course, to apply it to games, we'd have to get 100 game developers to all make their own version of some well-known game (like Minecraft) - then we could compare the results.