Monday 14 April 2014

Live TV Trials, Judges & Probability



In the midst of the Oscar Pistorius trial there has been some debate about whether live TV trials are a good thing or not. I suspect it’s as broad as it is long, but ironically the probability of whether live TV trials are a good thing or not is similar to the probability the judge faces in the trial itself (there are no juries in South African trials).

In both situations we have a probability of 1 or 0, but asymmetry of information too - so even though there is a decision to be made, its efficacy is based on a system where, when considering 1 or 0, all bets are made up of amateur decisions made by one person with incomplete information. If the judge thinks that Oscar Pistorius is guilty with a .6 probability of being right then that's as good as a guilty verdict. If the judge thinks he's guilty with a 0.3 probability of being right then that's as good as an innocent verdict.

In other words, all probabilities less than .5 are tantamount to one verdict, and all probabilities greater than .5 as are tantamount to the opposite verdict. That is why I say the jurisprudence mechanism is 'amateurish' - not that it necessarily needs to be revised (although I do offer some suggestions for revision in these Blog posts here and here) - but that it operates under a loose probability system that you'd never see in a science lab, or in manufacturing, or engineering, or cartography, or NASA, or places of a similar nature. For that reason, the case *for* live TV trials and the case *against* them may not have much in them in terms of their persuasiveness. 

* Photo courtesy of digital spy
/>