small percentages adds up scary fast over continents
'"Better that ten guilty persons escape than that one innocent suffer," says English jurist William Blackstone. The ratio 10:1 has become known as the "Blackstone ratio.'
It's not quantifiable. It's not a statement of social utility, and no utility function (e.g. which related community wealth to innocents jailed) could replace it. The maxim is a bedrock principle of how society defines justice. It's an articulation of the values and preferences that go into defining a utility function. That is, how just you _believe_ the government is a utility in and of itself, regardless of objective efficacy, and the maxim outlines one of the criteria for how you guage justness.
The notion is that a government that emphasizes protection of the innocent is believed to be most capable at defending fundamental liberties, in part (in as much as people rationalize it) because it incentivizes rigor in how government prosecutes criminals. Putting criminals away benefits everybody, so there's already substantial incentive to get it done. Imposing constraints on how it's done allows society to shape the process and, ideally, push it toward a global optimum rather than falling into a local optimum.
So it's an articulation that society places substantial value on liberty even if to the detriment of some security. (Compare that to cultures like Japan.) Second and simultaneously it articulates a belief that security _depends_ on the protection of liberty; that is, effective governments safeguard liberty, and those which do not safeguard liberty are unlikely to be effective at providing long-term security.
Neither of those statements are absolute truths, and they certainly don't depend on empirical efficacy. They're normative value judgments in and of themselves, therefore they help define the utility function, both in the aggregate and of individuals in the society, the vast majority of whom will have internalized those normative judgments.
YES, almost all doctors and judges will tell you 99% success rate is worth it. It takes a lot of work to educate them this paper 1% translates into 66% error rate in the field.
I expect it would be highly optimistic to even assume a person targeted by police is even three times as likely as the general population to be in possession of drugs, which would only be about 60%