Monte Carlo Simulation for list of Values

[KCG]

Hello all,

I am monitoring the performance of a machine. For each month listed below,
it broke down after running for the respective hours:

Month Hours
Jan-07 39
Feb-07 13
Mar-07 21
Apr-07 20
May-07 21
Jun-07 23
Jul-07 35
Aug-07 39
Sep-07 1
Oct-07 8
Nov 07 ?

For the month of November, using a Monte Carlo simulation (or similar) in
Excel, I want to predetermine the number hours that the machine will run
before it is next due for maintenance.

Thanx.

 

[Bernard Liengme]

How about =AVERAGE(other months) ± STDEV(other months)
But with such erratic data I would not bet my pay on the answer
best wishes

 

[joeu2004]

I question the value of a Monte Carlo simulation in this case. I hope
there is some discussion of that by some of the statistics experts
that participate in this newsgroup.

In any case, IMHO, before you can do any simulation or extrapolation
of data, I think you need to answer (for yourself) why there is such a
dramatic reduction of numbers in Sep and Oct, looking at it
"holistically", not just satistically. (Sigh, I cannot think of the
correct word to explain this.) That might influence the number for
Nov far greater than any analysis of the complete historical data.

PS: It is not clear to me what you mean when you say, for example,
that it "broke down after running for" 1 hour in Sep. Do you mean
that it failed repeatedly after 1 hour? Or do you mean that it broke
down once after running for 1 hour? If the latter, does that mean it
was ran for many hours before and/or after the breakdown? Or do you
mean that that the machine was no longer accessible for the rest of
the month after the breakdown; for example, out for repair? In
summary, does "1" represent a good bad or an extremely bad month?
These questions probably have little bearing on the answer that you
seek. I'm just curious.


----- original posting -----