I reverse engineered HLTV3 Rating Formula with +-0.03 error
Mix of Machine Learning and pure guess
RMSE 0.032, R^2 0.991
Used for MM scouting in SAURON
Main challenges:
- limited dataset
- swing itself is a hidden probabilistic rating feature
- only general idea published b4
some of my accounts / infrastructure were compromised. team is still mapping the full scope, so until i give the all-clear, treat anything from me with caution, especially links, dms, downloads, and anything urgent asking you to act or install something.
what i can confirm is safe: my email ([email protected]) and my core accounts (X, signal, discord, github, matrix).
how to verify it's actually me: identity + pgp key at https://t.co/fwqtCWRkdm
fingerprint: 1222 34C3 2EC5 21D0 626B 260D 094F DAEF 95F5 EDA6
signed statement below. anything that doesn't verify against that key isn't me.
lowk getting a bit burnt out w the automotive industry (as a career, will still be building cool stuff and posting about it) but if anyone has leads on entry/mid level dev positions or defense/adjacent manufacturing engineering/tech roles plssssss lmk :3