A note on odds
Posted by Roger Hunter on January 19, 2013 at 12:22:34:


Hi all;

Here's some information for you to ponder.

First, a table:


  M   QKS   DAYS  ODDS
  4 249217 14423 .9988
  5  59213 13937 .9652
  6   4939  3889 .2693
  7    568   539 .0373
  8     40    38 .0026
  9      2     2 .0001


The table is from my copy of the NEIC catalog
from 1973/01/01 to 2012/07/16, a total of 14440
days. (Yes, it needs updating)

M is magnitude, QKS is the number of quakes of
that size in the file, DAYS is the number of days
containing a quake of that magnitude and ODDS
is the probability that a day chosen at random
will contain one or more quakes of that size.

You will notice that the number of days is
never as large as the total day count. This is
due to clustering. There can be more than one quake
of a given size in a given day. So if you compute
the odds as number of quakes divided by number of
days, you'll be wrong.

That's why I compute odds as number of days (or windows)
containing quakes divided by number of days (or windows).

Roger

Follow Ups:
     ● Re: A note on odds - AMIT  11:59:50 - 1/21/2013  (100168)  (0)
     ● Re: A note on odds - AMIT  11:49:44 - 1/21/2013  (100167)  (1)
        ● Re: A note on odds - Roger Hunter  12:29:52 - 1/21/2013  (100169)  (0)