Lets first talk about Valve’s roadmap for Counter-Strike Global Offensive, mentioned briefly by Valve’s Gabe Newell in a reddit AMA he said the following:
As far as a roadmap is concerned, our priorities for 2017 are to replace the UI with Panorama, to make CS:GO available in more territories where a lot of Counter-Strike fans don’t have easy access to it (like China), and anti-cheat. Of course, we’re also planning on continuing to ship bug fixes and new features throughout the year, as in the past.
So clearly they are or from what Gabe has said looking at improving the game’s anti-cheat. Hard to believe as they haven’t exactly done anything significant in the past year or so but who knows we shall see what happens.
A user asked about why they don’t add automated detection for spinbotters and send them to Overwatch they mentioned that heuristics can be very dangerous as cheat developers could probe the system to determine the edges of the detection approach and that it opens up a lot of room for false positives. I personally agree with them but seriously they could do something if they really wanted to that would be adequate. Their focus however seems more on machine learning and this is reflected actually by the current deployment of VAC modules.
Looking at a very recent dump of the modules we can see that their timestamps date back to last year.
Now at first I thought they were faking these timestamps but it seems they are in fact using dated modules. Comparing the timestamps to ones that I’ve dumped on those exact days we can see there are no differences in function calls or anything really.
So for now we can predict that they are working more on a server-sided heuristics approach to detecting cheaters (like mentioned in the reddit post) instead of focusing more on detecting cheats themselves.
We shall see what happens.. 😮