Forecasting Help, Drag Down

Yes, it is "projecting" your data. Excel is assuming a simple linear relationship (y=mx+b where y is the data, x is count numbers ({1,2,3,...}). It regresses m and b from the data, then "projects" to the next count numbers (13,14,15...). Only you would know if this is a "good" projection or not, but that is how Excel is generating those numbers.

Originally Posted by coeeze

Is regression preferred over multiplying by a rate or moving average to generate a forecast? In my case the data is seasonal.

You are correct that this more math/stats than Excel. Linear regression like this would only be preferred if you expect the data to actually follow a simple linear trend. A simple linear trend cannot account for seasonality, so I would not expect linear regression to be preferred in this case.

Originally Posted by coeeze

do you know if there are any functions in Excel I could use to find m and b?

List of available statistical functions: https://support.office.com/en-us/art...__toc309306716 There is the =SLOPE() and =INTERCEPT() functions. I generally use the =LINEST() function. The =AVERAGE() function could be formulated to create a moving average, if you decide that is a better approach. Other statistical methods can be used as well.