Raul took what I wrote and ran with it in one direction but my only point
was that the 2nd version (34^~-.0.01) is nice because it directly
incorporates the two numbers mentioned rather than using (what we all know
is) 0.99=1-0.01 (or -.0.1).

There's a possible example of Bayesian updating here if, instead of a point
value like 0.007, we use a distribution centered on 0.007 and update this
estimate based on subsequent information like how many actually die of the
disease within a relevant period, say 2 weeks.  Introducing a time limit
like this also speaks to Raul's addition of a baseline mortality rate of
something like 1%/year which also needs to be considered in a more
comprehensive model.

On Thu, Oct 8, 2020 at 3:21 PM Brian Bambrough <[email protected]>
wrote:

> The chance that I survive is 0.99.  The chance that you survive is
> 0.99.  The chance for each of the other 32 to survive is 0.99.  The
> chance that we all survive is 0.99^34 or 0.71.
>
> On 10/8/20 12:44 AM, Devon McCormick wrote:
> > If 34 people have a disease which is fatal about 1% of the time, what is
> > the chance that no one in the group dies?
> > )
> > 2020 10 7 23 42 40.156
> >     0.99^34
> > 0.710553
> >     34^~-. 0.01  NB. Stating it another way
> > 0.710553
> >
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>


-- 

Devon McCormick, CFA

Quantitative Consultant
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to