Back in 2020, we went on a quest to model Aim stats more accurately than any other CS2 stat tracker (especially time-to-damage, crosshair placement, and spotted accuracy). This meant improving beyond the unreliable in-game radar events, and implementing our own raycaster to determine when someone is spotted on your screen.

However, it was still early days for us, and limitations in our raycasting tech meant that we used a very rudimentary approximation of a player to check when someone was spotted. In the image below, the 8 gray points at the corners of the orange prism were the previous approximation. The points on the person-shaped blobs are our new hitbox checks, which has improved the accuracy of our "enemy spotted" checks.

Old Hitbox (gray points making a box) vs New Hitbox (human shape)

The old method was honestly surprisingly effective given how approximate it is, but had a few scenarios where it led to inaccuracies:

  1. An enemy's hand is on your screen, poking around a wall (they are spotted in reality, but our raycaster would mark them as not spotted)
  2. An enemy is just barely behind a small piece of cover, but our old hitbox would poke out from that cover (this would mean we marked you as having spotted them before you did).
Example where you can see an enemy's hand

This new hitbox we're using has changed the overall averages for all Aim Stats slightly. Time-to-Damage and Crosshair Placement have both seen the biggest shift, getting better for many players on the new model. This means that lots of you will see a +4 change in your Aim Rating on your profile, for now.

Once we have a few weeks of data, we will re-calculate our benchmarks (the same way we did in 2024) to fit the new values that people are getting with the new system and re-normalize the average of Aim Rating back to 50.

As always, if you have questions or feedback, you can email us or reach out on Discord. We love hearing from you!