Re: one in eight
- From: "randovaro" <randovaro@xxxxxxxxx>
- Date: 15 Feb 2007 18:35:37 -0800
On Feb 16, 12:41 am, "p4o2" <p...@xxxxxxxxx> wrote:
David Winsemius wrote:
randovaro <randov...@xxxxxxxxx> wrote in
news:12t6qaf8771bc15@xxxxxxxxxxxxxxxxxx:
David Winsemius wrote:
Whoa, hold on there, partner.
Remind me... who said 12MM police-reported crashes? (The NTSA said
6MM in 2004. See citation above)
Explain again ... exactly how does 6MM get inflated to 34MM by GEICO?
And where are the NSC statistics that say "one in eight"?
I was just using the figure from your previous post:
"Crashes: The General Estimates System shows that an estimated
12,173,000 drivers were involved in police-reported crashes in 1996."
The NTSA figure I think refers to 6M crashes in total. This makes
intuitive sense if we assume that the overall majority of vehicle
accidents involve one car hitting another. Hence two drivers involved
for each crash. So 6M crashes=12M drivers in crashes. Double that to
24M crashes to take into account minor collisions (according to
"various sources") for the 183M or so drivers in total and you get
about 1 in 8 drivers involved in a car crash in the U.S. for that
year.
Got it. A million here, a million there, pretty soon you get real risk.
More recent information on recent licensure numbers at:
http://www.fhwa.dot.gov/policy/ohim/hs05/driver_licensing.htm
2004 had 199MM licensed drivers and if you add up the numbers of drivers in
crashes from table 21 of the earlier offered cite:
http://www-nrd.nhtsa.dot.gov/pdf/nrd-30/NCSA/TSFAnn/TSF2004.pdf
... its only 10.3MM drivers. Even with the doubling, we only get to "one
in ten".
Your hypothesis describing how someone at GEICO with a phone call to the
National Safety Council might have cobbled together "one in eight"
estimates from disparate sources in an unspecified time period using
adjustments from "various sources" remains unchallenged. I withdraw my
reservations.
--
David Winsemius
You have been great putting meaning to the "one in eight". So if
Allstate would add "in the next year" to the commercial it would be
OK. In the references is there a way to include the drivers that have
2 or more in a year? For instance a driver has a one in eight chance
of having an accident. Like a coin toss, does he (she) now have a one
in eight chance for a second?
Well it appears Google Groups has lost my original reply. I'm trying
again so apologies if this appears as a double post.
What Allstate should say is that there's a one in ten (may as well
update the data) chance of you being involved in an accident (incl.
very minor collisions) *per year of driving*. The chance of a
collision isn't 1/10 every time you get behind the wheel. Nobody
would go outdoors or premiums would skyrocket. There's a time factor
to take into consideration. So that's 1/10 during a year of driving,
1/3650 during a day, 1/87600 during an hour behind the wheel (on
average), etc. Over a typical person's entire lifespan they would
probably spend maybe 5 years or less actually driving? Assuming that,
and using the binomial distribution, the chance of you being a driver
involved in at least one collision at some point in your life (in the
U.S.) is therefore:
1-B(0,5;0.1)=0.40951
All things being equal, a driver's chances of being in an accident is
independent of their history, like a fair coin toss.
.
- Follow-Ups:
- Re: one in eight
- From: David Winsemius
- Re: one in eight
- References:
- one in eight
- From: p4o2
- Re: one in eight
- From: randovaro
- Re: one in eight
- From: David Winsemius
- Re: one in eight
- From: randovaro
- Re: one in eight
- From: David Winsemius
- Re: one in eight
- From: randovaro
- Re: one in eight
- From: David Winsemius
- Re: one in eight
- From: randovaro
- Re: one in eight
- From: David Winsemius
- Re: one in eight
- From: p4o2
- one in eight
- Prev by Date: Re: Testing Variance´s Homogeneity
- Next by Date: Re: cramer-von-mises
- Previous by thread: Re: one in eight
- Next by thread: Re: one in eight
- Index(es):
Relevant Pages
|
