Here's why I hate games as applications for the typical one-man software development shop:
1) They require a huge upfront investment in asset creation.
2) The assets start depreciating instantly, because gamer expectations are constantly advancing. (Expectations for all apps advance, but for games it is particularly acute. In five years Bingo Card Creator will still look like Bingo Card Creator, but a 2013 game which resembles a 2008 game will be virtually unsaleable. Look at how much even "casual" games have advanced in the last 24 months if you don't believe me.)
3) 1+2 means that the post-launch sales curve is decreasing, rather than increasing (as it is for most apps). This makes post-launch marketing and other improvements largely a black hole of effort, instead of a series of steps one can proceed through to build value.
4) Games are typically needs not wants. Accordingly, it is both hard to convince people to pay for them and hard to market them via search engines, because people don't typically know they want a [match three game with dragon theme and some differentiating feature].
5) Hit driven -- winners win, everyone else takes a bath on asset development costs.
6) The people who play them most are those who are least able or willing to pay for them, and most capable of getting them for free
The App Store brings the fun that is writing a game to any other field of software development, with the exception of point #6. Selling a game on the App Store brings you to the double whammy -- you have constantly depreciating art assets AND the fall-off-front-page-and-watch-sales-die effect contributing to murder your post-launch sales graph.