[dragonwriter]

> It's just one datapoint, but under 5% of the thefts this specific taskforce identified were perpretrated by organized groups.

5% of the people arrested, not 5% of the thefts.

“Less sophisticated thieves are more likely to be arrested” isn’t a surprising result.

[pixl97]

Yep, much harder to get caught if you have a team monitoring for police activity and a ready get away car and they know where to hit with the highest response times.

[gorlilla]

It's still accurate in the sense of proximate cause. Each theft event is a singular event composed of one or possibly multiple items being stolen.

[dragonwriter]

> It's still accurate in the sense of proximate cause

Its not, though.

> Each theft event is a singular event composed of one or possibly multiple items being stolen.

Even if you ignore that the arrest rate likely isn't consistent between the two groups, that doesn't make the arrest proportion “accurate in the sense of oroxinate cause” unless you assume also that the organized vs. disorganized arrested parties are responsible for the same number of theft events per individual.

All the proportion of individuals arrested in the theft ring vs. not actually tells you is the proportion if individuals areested. It doesn't tell you the proportion of “theft events” or the proportion of $ valie of theft or anything else.