The aim of this repository is to provide the public (n=50) sample of server-side reach based off different latencies (0ms, 120ms, 200ms, 400ms).
All of the collected data is in the samples.go file.
The example code used to collect the data is in example.go
Furthermore, the reach data collection was done by using a 10cps autoclicker to put a bot player in a combo for 50 sample clicks.
The data was collected and then put into a basic one-sample t-test with the null hypothesis being that the reach was equal (or less than) 3.00 while the alternative hypothesis was that reach exceeded 3.00
The results found no statistical significance (SS) evidence that reach exceeded 3.00 for the legitimate client, while there was SS evidence suggesting that the 120ms, 200ms, and 400ms groups had higher server-side reach.
In simple terms, there is evidence to show that higher latencies generally have higher server-side reach
Basic flaws with this study were differences in rotation and position of player per knockback which made reach sometimes be legit or not legit. Of course we could always have larger samples and this was just an extremely basic stat. test.
Credits:
- Response to The Misconception of Elevated Latency Advantage by EL1SP.
- T14 Raptor for providing the ray tracing code to allow for server-side reach calculation and his input on the document.