Joshua Halpern wrote:
wrote:
Roger Coppock wrote:
"~80,000 return period has a 95% confidence interval of ~8000- 800,000"
Thank you Harold for the 95% error bars.
Yes, I thank Harold too, for having more patience that I for
explaining the details of your statistical mistakes.
(That is why I used
the word "About" in the title.)
Using "about" 80,0000 for a range of 8000 to 800,000 is as bad
On a log scale it is only 2 and what is a factor of 2 among friends?
josh halpern
[chuckle] Quite a bit, as you know. When I was in grad school my
research group calculated a lot of power spectra which when plotted
as semilog or log-log plots gave approximately straight lines over
several orders of magnitude, but deviations on the plots which were
almost unnoticeable represented large real deviations. One time our
prof needed to point this out in a paper, so in the caption he wrote
"Note the log-log plot". Thereafter the running joke whenever one of
us showed a result at group meetings, etc., plotted that way was to
remind everyone to "Note the log-log plot". Hmm, I suppose you had
to be there...
Cheers,
Russell