The main issue is that this particular dataset (voting results from precincts of very similar size) is not a valid candidate for BL analysis.
Financial data, which will have values over multiple orders of magnitude, is an excellent candidate. But even so, it's not proof. However, it can be a powerful indicator to justify a deeper dive.
Continuing to promote BL for precinct-level data just makes us look dumb and provides clickbait to the left.
Considering I actually have forensic accounting in my background.... no, that's not how it works.
You use anomalies to dig further and find proof. You have something like Benford's that highlights a problem. "Hey, look at this, something screwy". Then you start digging to find the anomalies, and audit from there. Audits are where the discovery comes from, and that is what allows you to find proof.
The key difference here is that, as said before, y'all are grasping at straws saying "LOOK AT THIS PROOF". This isn't proof, it is highlighting the need to do a deep dive and FIND proof. When you present data analysis alone as concrete evidence and it is then (expectedly, I might add) tossed in the garbage at court, y'all get all surprised Pikachu face, even though literally every lawyer with a podcast and YouTube channel knew it was coming.
Couldn't they have leveraged these analyses as probable cause to get subpoenas for further investigation? Always puzzled me why they didn't. Asking for the moon with just "data analyses" that haven't even been cross-examined was always going to go down in flames.
It's utter stupidity to think an anomaly is proof positive enough for court. Insane. Anomalies are and always will be just enough to give cause to further investigate.
Law enforcement get warrants issued based on anomalies... they wouldn't just immediately charge the suspect and go to trial with just that base anomaly (unless they suck at their job).
If this were true then there wouldn't be people sitting in white collar prison for financial crimes you fucking moron.
Except anomalies are cause for further investigation, not the actual end proof of something nefarious -_-
The main issue is that this particular dataset (voting results from precincts of very similar size) is not a valid candidate for BL analysis. Financial data, which will have values over multiple orders of magnitude, is an excellent candidate. But even so, it's not proof. However, it can be a powerful indicator to justify a deeper dive.
Continuing to promote BL for precinct-level data just makes us look dumb and provides clickbait to the left.
Considering I actually have forensic accounting in my background.... no, that's not how it works.
You use anomalies to dig further and find proof. You have something like Benford's that highlights a problem. "Hey, look at this, something screwy". Then you start digging to find the anomalies, and audit from there. Audits are where the discovery comes from, and that is what allows you to find proof.
The key difference here is that, as said before, y'all are grasping at straws saying "LOOK AT THIS PROOF". This isn't proof, it is highlighting the need to do a deep dive and FIND proof. When you present data analysis alone as concrete evidence and it is then (expectedly, I might add) tossed in the garbage at court, y'all get all surprised Pikachu face, even though literally every lawyer with a podcast and YouTube channel knew it was coming.
Couldn't they have leveraged these analyses as probable cause to get subpoenas for further investigation? Always puzzled me why they didn't. Asking for the moon with just "data analyses" that haven't even been cross-examined was always going to go down in flames.
It's tough when the courts throw your case out without ever examining your evidence.
Get better lawyers that won't fumble procedural issues, and that don't sue in the wrong court.
You're getting downvotes for being right.
It's utter stupidity to think an anomaly is proof positive enough for court. Insane. Anomalies are and always will be just enough to give cause to further investigate.
Law enforcement get warrants issued based on anomalies... they wouldn't just immediately charge the suspect and go to trial with just that base anomaly (unless they suck at their job).