It seems that many people are using standard spreadsheets to forecast, getting correlation coefficients, etc. However I decided I needed to write my own computer program so I could work out the likelyhood of the forecasting being correct. 1) If there is not sufficient correlation then forecasting is a dubious business 2) If you do find sufficient correlation, what is the 95% interval for your forecast.
I then thought that if people wanted me to help them with these two things it would be better if they did not have to send me long lists of data, so I now have a program that works if you send me such things as the sum of (the x-values squared). It tells you whether there is sufficient correlation in the first place using a significance table for the correlation coefficient and then gives a 95% confidence interval for the forecasted value. Does the average person using a spreadsheet know about this sort of thing? If anyone wants me to send them the program free I can do so. It is an exe program of about 200KB (small) and if you have a Yahoo email address I could email it to you (Gmail, etc do not allow exe files to be emailed). Was completed today and does not seem to have bugs.