GeorgeRaetz
Jan26-11, 11:40 AM
11 US police officers were murdered in line of duty over a 24 hour period (www.foxnews.com/us/2011/01/24/authorities-fear-cops-targeted-officers-shot-hours/#content)starting about 23Jan2010. Is this a statistical fluke, or not? Here is my calculation:
In 2010 there were about 90 non traffic officer deaths ,mostly shootings. That's a rate of lambda=90/yr
Assuming a random distribution over time, the probability of k deaths after a time t with death rate lambda is given by the Poisson formula:
P(t,k,lambda)=exp(-lambda t) (lambda t)k/k!
Plug in values t=yr/365, k=11 I get:
P(yr/365,11,90/yr)~4 10-15.
In other words, the probability that 11 cops would be murdered in a one day period is about one over 2 1014.
Given that there have been far less than 1014 days since the Universe began, that does seem to be an extraordinarily unusual day.
Or have I made a mistake in calculation or logic?
In 2010 there were about 90 non traffic officer deaths ,mostly shootings. That's a rate of lambda=90/yr
Assuming a random distribution over time, the probability of k deaths after a time t with death rate lambda is given by the Poisson formula:
P(t,k,lambda)=exp(-lambda t) (lambda t)k/k!
Plug in values t=yr/365, k=11 I get:
P(yr/365,11,90/yr)~4 10-15.
In other words, the probability that 11 cops would be murdered in a one day period is about one over 2 1014.
Given that there have been far less than 1014 days since the Universe began, that does seem to be an extraordinarily unusual day.
Or have I made a mistake in calculation or logic?