Hacker News new | past | comments | ask | show | jobs | submit login

at some point, statistical proof is proof.

that's how DNA evidence works, after all.




No, it really isn't. The DNA is the evidence, the stats are used to bolster the weight of that evidence. Stats are just numbers, they are not proof. Nobody ever got convicted on 'just stats' as proof, it will always be in the context of something else.

Just saying something is extremely unlikely to happen by itself does not constitute proof, after all it just could be that that extremely unlikely thing happened. You'd need to corroborate that with something else to meet the burden of proof that is normally used in a criminal case.

Different legal systems have different standards of evidence before being allowed to consider a defendant proven guilty, as far as I'm aware nobody ever got convicted on math alone. There always was something else.


> Just saying something is extremely unlikely to happen by itself does not constitute proof

sure it does.

statistics are a measurement of something in the real world. in the poker example you are measuring the possibility of a series of hands succeeding in a given episode of play. In court you are measuring the possibility of a given DNA sequence occurring by chance.

What distinction are you drawing between the statistical certainty in poker vs. the statistical certainty in court (or in medicine)?

There doesn't need to be anything else, if you have a sufficient statistical level of certainty.

In most cases there is corroborative evidence, but nobody thinks it's necessary when the DNA matches.


> Just saying something is extremely unlikely to happen by itself does not constitute proof

That is literally what it constitutes.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: