Worldwide, there are 17 earthquakes
per year of magnitude 7 or above. But
in 2010 there were 24 of them. Why?
Was Nature especially angry in 2010?
In 2012 the total fell back to 14. The
result of some top-secret anti-
earthquake device?
Of course, the ‘17 per year’ is an
average figure. From 2000 to 2012
the annual total varied from 11 to- It’s always tempting to look for
causes, but sometimes what we can
see is the noise, not the signal: natural
variation around a steady average.
When numbers are small, this variation
will seem more extreme, like our
earthquake example.
Mary Huntrodds found a husband with
the same birthday as her. The chances
of which are small, as we’ve seen. But
Prof Spiegelhalter only noticed their
memorial because of their shared
dates. So perhaps we should be asking
a different question: is chance
subjective? What are the chances of
somebody who catches the eye of a
Professor of the Public Understanding of
Risk being involved in a coincidence? It
would be more surprising if he’d
noticed a memorial with absolutely
nothing unusual about it.
Looking backwards, it’s easy to say
what proportion of couples share a
birthday, and calculate the odds thatAnother example is speed cameras.
Because they’re usually installed after a
high number of accidents in one place,
it’s likely that the number of accidents
would fall the next year even if you didn’t
put up a speed camera. Because of this
regression to the mean, the effects of
speed
cameras can
look stronger
than they
really are.Random events come
along at regular intervalsIt’s easy to predict the
future from the pastany one couple did so.
But looking forwards is
trickier. Are you confident
future couples will
behave like past
couples? Will you only
look at people who ended
up married, or should you
also include single people
who don’t go on to
become couples, in case
September babies are
especially lucky in love
and skew your
figures? The questions
you ask will affect the
answers you get.The plaque marking the graves
of Mary and Francis Huntrodds
dddidididd s s s so.o.o.o.
wwwwarararadsdsdds iiis ss s
cococonfnfnfnididdenenenett t
wwwililill l
ttt
u uu ononononlylylyly
wwhohohoo eeendndndndededed
hohohohouluululd ddyoyoyyoou uu u
glgge eeeepepepepeopooplele
tototo
sss, ,ininn c ccasasse e e
eseses aarere
iinn lolooveveeeeeststioionsnsn
ccct tthththhe e
t.t.t.t.