Thanks Frank, it is great to be taught by the best. This is the strength of the SML, it connects those asking questions, eager to learn, with the best teachers. Often the roles reverse. We all benefit.

Regards Roger

--------------------------------------------------
From: "Frank King" <[email protected]>
Sent: Tuesday, February 15, 2011 9:31 AM
To: "John Carmichael" <[email protected]>
Cc: <[email protected]>
Subject: Re: part 2 of longitude correction

Dear John,

I sympathise...

I flunked statistics in college.

I never got the hang of statistics until I
was fingered to give a lecture course on
the subject and I have never looked back!

In dialling, I use statistics for error
analysis and also when trying to do things
which are impossible!  Rotating a horizontal
dial for Summer Time illustrates the kind of
thing.  You can use a best-fit approach to
arrange the hour-lines so they approximately
work in both states.  Hey, I am still on the
topic!

I am now in the middle of planning a talk at
the April BSS Conference where I propose to
describe a quite different kind of impossible
sundial!  Come along!  No statistical ability
needed :-)

Your analysis of the Equation of Time is not
quite as bad as some readers might think...

  Best case:  off by 0 minutes
  Worst case: off by 16 minutes
  Average:    8 minutes.

The result, 8, is known as the Mid-Range value
and is related to the Mean, Median and the rest.

The mid-range value is generally regarded as
a bad choice because it is governed by two
"outliers", the extremes of 0 and 16.

With random numbers such use of outliers is
indeed bad but with the Equation of Time we
are not using random numbers and the result
is not too far from more respectable values!

In the opening lecture of my course I always
issue a warning: "Never accept an invitation
to play a betting game with me!"  Be careful!

All the best

Frank

---------------------------------------------------
https://lists.uni-koeln.de/mailman/listinfo/sundial

---------------------------------------------------
https://lists.uni-koeln.de/mailman/listinfo/sundial

Reply via email to