But for the moment other markets also exist, such as desktop where one can still be very successful (as long as you can hit a certain level of word of mouth).
For a niche product each happy user would have to tell many other potential user to make one conversion. So practically niche products rarely get viral.
Hence why Steam focuses on games that can be sold easily, not especially those with the happiest users. To make a game greenlit the voters vote on the screenshots and the beginning of the video, not the gameplay.
Game development is more of a passion / calling / addiction than a rational business decision. It's an even dumber idea than startups. But if it's all you want to do, then...
Now, evidence is a hard word, but I've the hunch that unless your game is an addictive style freemium incremental, all other genres are as of today more suited to desktop, because of user disinterest on paying a mobile game upfront.
It's far more difficult to gain recognition and get word of mouth going on the app store (due to volume), combined with the lower selling point as compared to desktop (for example $1 vs $15). People don't want to pay up front for games and the race to the bottom on prices have really hurt the little guy.
All graphs I've seen of sales on the app store show an inordinate amount of sales for the top sellers, whereas the tiers in the middle and below don't make anything significant at all.
Then go look at steamspy etc stats, even games 2/3rd's down the lists for a year's releases would have made a few thousand $ (as compared to app store where they might have been sitting on tens of $).
Either way, learning to code via writing a game is something I recommend for anyone, anywhere.