We often talk about the science that are free-to-play monetization systems, and this story definitely shows how there's no stopping this transformation. Scientists have developed a model called "survival ensemble" that can predict with pretty good accuracy on what day and at what stage of a game a user will stop playing and why. This new algorithm can predict players' reactions with pretty good accuracy. By using this group of algorithms (around 1000 submodels all looking at different variables and conditions), the developers can get a good idea of how their game will perform and can change it accordingly to ensure players stick around longer.
Apparently, these kinds of models are used in medical research and biology, but now they are being used increasingly in mobile games. And, as you can imagine, they aren't being used by small developers but, rather, by those who can afford such tools. Curious to see how free-to-play games will look in a few years if developers continue sharpening these tools and changing their games accordingly.
[via MoneyControl]

...