Depends on if the toolset you're using makes it relatively easy to port your game to Linux. If it's easy, it's worth doing. Or even if it's not that easy, but your game is making a lot of sales on Windows and Mac, then it's probably still worth it. But if it's a lot of work and has low potential, don't bother. It's all about cost/benefit analysis. Is the potential advantage of having your game on Linux, worth the work involved? The answer depends on the specifics of your project.
There are a lot of Linux systems out there overall, sure, but Linux only has a modest share of the desktop category. By most measures, perhaps 2-3%.
Mac, maybe 6-8%, Windows the other 90%. So as far as higher-end desktop systems go, Windows is still persistently dominant.
But factor in all the tablets and phones and the fact that Android derives a lot from Linux, and the sheer number of digital cameras and little 'smart devices' that run off Linux variants, and you'll realize Linux is pretty much everywhere in one form or another. There are hundreds of millions of devices running off Linux.
Note: The most widespread OS now is Android. There are over a billion phones and tablets using it. But Android users do tend to be stingier when buying things; the average Android user spends way less on apps than the typical iOS user.
After that, Windows and iOS, are the next two biggest markets. Each of these markets has in excess of 400 million devices active.
Then Mac OS X and various forms of Linux.
My project 'Miniature Multiverse' is currently being built with Unity and my aim initially is for desktop Windows/Mac/Linux but I'm also hoping to launch a mobile version perhaps, some time after, if things are going well. With Unity my game is not too much work to port across desktop systems. The only reason I'm slower on mobile, is that I'd need to figure out the touchscreen interactions before that'd work. But I'd love to have my game running on mobile!