[time-nuts] Rubidium standard

Steve Rooke sar10538 at gmail.com
Wed Nov 18 08:18:52 UTC 2009


2009/11/13 Mike S <mikes at flatsurface.com>:
> I'm sure someone with more statistics background can add to this, but useful
> (or expected) lifetime cannot be determined from an MTBF number.
>
> Here's an example I found, demonstrating this:
>
> "There are 500,000 25-year-old humans in the sample population.
> Over the course of a year, data is collected on failures (deaths) for this
> population.
> The operational life of the population is 500,000 x 1 year = 500,000 people
> years.
> Throughout the year, 625 people failed (died).
> The failure rate is 625 failures / 500,000 people years = 0.125% / year.
> The MTBF is the inverse of failure rate or 1 / 0.00125 = 800 years.
> So, even though 25-year-old humans have high MTBF values, their life
> expectancy
> (service life) is much shorter and does not correlate."

I don't know where you found this statement but I's check and change
you get from them for wooden nickels. Humans have predefined lifetimes
and such calculations for MTBF does not apply here. The failure rate
of a human is not constant over the lifetime and just taking a figure
at the age of 25 will get you nowhere. Try doing the same exercise at
ages 70 or 80 and you can see how it does not work. This is another
example of how statistics can be abused.

Very amusing though, thanks for posting this.

Steve

-- 
Steve Rooke - ZL3TUV & G8KVD
A man with one clock knows what time it is;
A man with two clocks is never quite sure.



More information about the time-nuts mailing list