The real issue is the cost of false positive detection of cheating is negligible since the vast majority of positives are probably true positives—it’s the cost of doing anti-cheat business (minimal)
But yes cheats would be modified to just below thresholds of detection
I think this might be in reply to my first comment about scaling? If so, I just want to clarify that I was thinking more along the lines of scaling the customer service/ban appeal side rather than infrastructure.
If, for example, every ban had a component of someone at Valve taking 10 seconds to review in-game stats at the time of the ban, and then making a determination of whether or not those stats seem reasonably non-cheater-ish (pretty hard policy question in itself), the process would slow to a crawl.
Sure - looking at K/D, accuracy, etc., is an important factor in a statistical model.
Statistics can also include: Map name, player transform on the map, keyboard and mouse events, GPU utilization, audio playback events, etc. These are all very high information time domain signals that can be correlated with the same from any other player.
After a certain point, I don't think it matters if it is publicly known what your signals are. The amount of information becomes overwhelming in aggregate. You can impose the curse of dimensionality on the cheater.
I don't think these are the type of stats the parent was referring to when they said "If someone at Valve had taken 10 seconds to review my stats".
But sure, those are all examples of statistics to start logging, analyzing, and cross-referencing. (I would argue most of the statistics you listed are of little to no use in identifying false-positives (or good cheaters), but I understand the point you're making with those examples.)
It would maybe reduce the false positive rate by some amount at an increased monetary (and complexity) cost to themselves. I think it would be well past the point of diminishing returns though. Setting up all the infrastructure, policy, processes just to reduce false-positive rates by a few percent, maybe?
I think I'll stand by "that's unreasonable" and "cheaters will game the statistics".
It seems super reasonable when it's a one-off thing for your own account. When you think about making it into policy and scaling it up to 1000s of interactions, it quickly becomes unreasonable.
>Statistical techniques can dramatically reduce false positives
For a period of time, anyways. Until the statistics get gamed by the cheaters (e.g. adjust accuracy of your auto-shoot from 100% to 85% or whatever).