aren’t they just confusing likelihood with posterior probabilities

https://www.abc.net.au/news/2025-12-02/statistically-predict-domestic-violence-homicide-risk-rare/106073508