Six Game Data Mining Myths

As game data mining is still something that is being defined, there are a lot of myths circulating about what telemetry data from games is, how it is used and what data mining of these data can do for developers and publishers. Read on for a bit of game telemetry mythbustin´.

Myths about game data mining

1- With game data mining, we can fire our designers – the testers/users will tell us what they want! Wrong: mining gameplay telemetry data is incredibly useful for evaluating and testing design, but telemetry data cannot tell you how your players feel or if they have a good experience – game data mining is not a replacement for good game design.

2- With game data mining, we do not need user testing! We can fire our testing department and save heaps money! Wrong: Game data mining goes hand in hand with user-oriented testing and -research, but does not replace it. With mining of gameplay metrics data, a powerful supplement to playtesting and usability testing is gained.

3- Game data mining is autonomous, requiring little human oversight! The tools are automated we just turn them loose on our data and find the answer to our design/business/marketing problem! Wrong: There are no automatic tools that will solve your problems – data mining is a process, as highlighted by CRISP. Additionally, data mining requires significant human interactivity at each phase, and also for the subsequent quality monitoring.

4- Game data mining pays for itself quickly! Lets invest in tools, infrastructure and people right away and save heaps of money! Not exactly wrong: The return rates vary, depending on the specific situation, the game, the size of the developer or publisher, etc. The return will be there in terms of improved knowledge, but careful planning needs to go into deciding on an initial setup and the strategic and practical goals.

5 – Game data mining will identify the causes of all our problems! We will make heaps of money integrating game data mining in our business! Not exactly wrong: The knowledge discovery process will help identify and uncover patterns of behavior in the data whether user-derived or business-derived, and these can be highly valuable, but it requires human interpretation to identify the causes of the patterns (with the help of analysis).

6- We need to obtain data on everything! Data equals value we will make a heap of money! Well…: You need the right data, to solve the problems you have. Just measuring everything will waste resource. Getting the right data requires as much thought as their analysis – and use samples of data rather than full datasets whenever possible.