That's the point the article is missing. Many people currently creating HTML5 games are not motivated by money. They can be hobbyists or people familiar with JS who wants to try making a game using a language they already know. As a consequence this lead to the apparition of many game engine/frameworks/libraries just like the last few years have seen an explosion of "generalist" JS frameworks/libraries. And this explains why the last few weeks we saw many flappy birds and 2048 clones. Well, and the Flappy Game Jam probably helped regarding the formers too.
And as these types of games can be quite addictive, while using JS made them easy to share over Twitter or Facebook, no wonder they were so successful.
For the moment there is nothing indicating that these two free successes could lead to the commercial success of paid web games. At least nothing so far that would make developers migrate from native apps.
