Re: trendline gaussian

Dave,

:eureka:

A light goes on. The key thing about curves that are used for probability distributions is that the integral of the function, that gives the area under the graph, totals 1. Now, if we knew how much oil there was in a given reserve we could then fit a curve to the area as well as the production points.

I have amended the worksheet to include all of the oil production data. It shows that it is difficult to get a Gamma dist to fit, but the normal curve is a very good fit!

I am not sure what you did to extend the references. The easiest way is to insert a 100 rows between, say, the 20th and 21st row of data and then paste the new values in and delete the extra rows. Then drag the formulae down. This way you do not need to amend any formulae that are dependent on the new data and the graph series will automatically extend.

Do you have lots of data sets that you wish to 'fit' the curve for? How many? If you do have a lot then we need to find a better way of doing it, however if you have say 10-20 then this method will work. One that that we may want to do is to 'window' the data and only select data for certain years.

HTH,

Alan.