Strange that it has come to this isn't it?
Gaming has always been an investment, and early game consoles and games were relatively more expensive that they are today (see:
http://arstechnica.com/gaming/2010/10/an-inconvenient-truth-game-prices-have-come-down-with-time/ or
http://www.1up.com/news/90s-game-price-comparison-charticle) and the truth is well developed games at reasonable prices simply aren't as safe an investment as a develop as you go/DLC-centric product. If you spend three years developing a title, and after the second year of development realize that play-testers and focus groups aren't responding well, or you've reached the realization that there is some technical (or other) limitation you can't overcome, you've lost at least those two years of development with no releasable product. If instead you try to cram the general mechanics into the game as rapidly as possible, and get the community play testing for you, and then monetize them to support continued development, not only do you take significantly less of an initial investment risk, but you also stand to profit off of your product for significantly longer than just the title's couple weeks or month at the top of the release charts.
TF2 is of course the king of this, and the long lifespan of it probably reflects valve's business acumen. If I was a game developer, and you told me I could not only sell my title, but then, as sales began to drop of re-monetize the user-base through micro-payments literally YEARS after the game's release, why wouldn't I say yes? Especially after the considerable (and probably costly) development that went into the original release (many years. and more than one total overhaul), and the subsequent updates before you could buy things for the game. The Mann-conomy update (introducing micro-payments) went live on September 30th, 2010 but there had been 11 major content updates, and two addition community content updates since TF2's release nearly 3 years earlier (October 10th 2007).