I have to admit the game is getting less time at the moment because I am flat out with my work on the SDK for Philips Hue. Now, Philips Hue is a successful project, great technology, great software.
I have done a lot of business software over many years. Most products have come on to the market, my average is good for software development. However, that is not true of the industry in general. It is reported in studies that less than 10% of software projects come in on time and on budget, and 70% of projects fail. The bigger the project, the more likely the failure, and I have seen that in practice, with one of the worst ones involving over 100 main developers.
So keep it small, keep it mean, keep it lean is best.