forecasting remaining months of seasonal sales - with four years of history

[Meredith]

I have four years of monthly sales numbers.
and 9 months of this years sales. I want to project the remaining months of sales.

right now i am using a combo of linest and intercept to project each remaining month,
LINEST($C56:K56,$C53:K53)*L53+INTERCEPT($C56:K56,$C53:K53)

meaning:
LINEST(current year monthly sales,prior year monthly sales up till prior month)*last year current month sales+INTERCEPT(current year monthly sales,prior year monthly sales up till prior month)


but i can only do that based on the prior year. not on all of the years.

Thoughts? I could probably do a weighted average of prior years weighting more recent years more to combine them all into one data set.