I talked some days ago with Ichiro from Dejobaan. They went with Aaaa! for PC on Steam and it was quiet a hit and made a fortune for them. The port for the iPhone / iPad was done by a 3rd party for which did revenue share and it sold some numbers, but not as much as expected for a 2nd launch buzz. He reported as well that Android games, if somehow successful, are reported to sell roughly only ~20% of the same revenue from the iOS version, if the iOS game is successful.

I feel that the market is broading and diversifying, but generally speaking I think there will be at least two types of games:

- fat games for desktops and Android based consoles based on mouse/keyboard and/or gamepad input with a strong internet connection
- thin games for mobile devices based on touch input with an unreliable, weak internet connection

If you do good PR prior the launch of your game, you can get a feeling if it will become a hit. If you have run a market analysis and found out which platform could generate massive numbers, hiring someone to do a port for a second launch buzz could work and would be more riskless than doing it by yourself and/or risking to do it in the first place.

So, developing for PC (classic fat game) and developing for Android (thin touch game) seems to me a riskless and valid option for multiplatform development. A later port to iOS, Windows Mobile, Mac or Linux can be considered later. I know only a few companies that do a real multiplatform launch at the same time for a game, so why should you do that?