The main pitfall of doing any "cost/benefit analysis" on Linux gaming is that most studios look at the short-term, "here and now" numbers and judge it from there. But there are a lot of intangibles that come with bringing your game to a free and open platform with a bunch of technologists and social media magnates just hungering for your games. If you just look at it in terms of immediate profits, most games based heavily on Windows-only technologies are not worth the cost.
Supporting Linux has to be a long-term strategy for the company, and it has to involve going "all in" on the Linux platform: supporting it as the premiere platform for all of your titles, and driving adoption of the platform by way of your games.
A lot of developers are rightfully fed up with the lack of flexibility and tyrannical behavior of Microsoft, and would benefit hugely from using a free and open operating system with no strings attached. But all they're willing to do is run "market studies" to determine the approximate number of people currently on that platform who are willing to buy their games, so almost always those calculations come up short. They're just taking the wrong approach. You can't just bring a game to Linux as casually as you might add PS3 support to a PC/Xbox game, because on well-established platforms you can just drop the game into retail stores and the profit will come with minimal effort on your part.
Most companies don't want to build a platform user base simultaneously with porting games to that platform. Valve is the thought leader in that space among game developers. Others need to follow in order for the movement to succeed, though, and put a forceful end to the notion that you have to keep Windows around to play games.