Environmental Economics

Using data from almost 9 years of pageviews at this blog, I fit a trend line with an OLS regression model where the dependent variable is the 30 day moving average (the blue line) of page views as measured by Typepad (where x is the time period). Here are two contrasting 1000 day forecasts (the trend line beyond the blue). First, the R² maximizing cubic function (the linear model has an R² of near zero):

Second, here is the quadratic:

Which forecast would you choose? I'm hoping for the quadratic because the light is at the end of the tunnel.