Why is my custom proxy detection method failing to identify specific residential VPNs?

we're seeing persistent false negatives with our custom proxy detection, specifically against residential VPNs. despite advanced fingerprinting and heuristic analysis, a lot of this traffic just bypasses detection, lookin like legitimate residential IPs. what methods beyond standard header analysis and known blacklist lookups are effective for accurate proxy detection against these highly evasive services, or am i missin something fundamental?