[Avodah] Math behind kulan chayav

Micha Berger via Avodah avodah at lists.aishdas.org
Sat Jul 29 20:12:50 PDT 2017


"Too good to be true: when overwhelming evidence fails to convince."
Lachlan J. Gunn, et al.
Proceedings of The Royal Society A. To be published
Paper: http://arxiv.org/abs/1601.00900
Magazine article on said paper (H/T R/Dr Josh Backon):
    https://m.phys.org/news/2016-01-evidence-bad.html


This paper focuses on the case of identifying a criminal in a police
line-up, archeological evidence, and crypography testing, but I'll just
use the first as an example. Sometimes a line-up is like picking out
an apple out of a group of bananas. But usually, there is measurable
probability of error. Whether it's 48% for a criminal who runs through
the crime scene or the probability of mistaking one's kidnapper, there
is always some probability of error.

Now, what's the probability that no one makes that error? If 10 people
unanimously identify the same person, is it more likely that no one
made the rror, or that there is a systemic bias leading people to
select the same member of the line-up?

This paper shows the math (and has plots for various probabilities of
error). It takes far fewer people than you'd think before a unanimous
finding becomes suspicious.

And the paper even cites the din as precedent:
    However, this is not necessarily the case and more confirmations can
    surprisingly disimprove our confidence that the defendant has been
    correctly identified as the perpetrator. This type of possibility was
    recognised intuitively in ancient times. Under ancient Jewish law
    [13], one could not be unanimously convicted of a capital crime --
    it was held that the absence of even one dissenting opinion among
    the judges indicated that there must remain some form of undiscovered
    exculpatory evidence.

-Micha

-- 
Micha Berger             Zion will be redeemed through justice,
micha at aishdas.org        and her returnees, through righteousness.
http://www.aishdas.org
Fax: (270) 514-1507



More information about the Avodah mailing list