Plotting overlapping prediction intervals

I often see figures with two sets of prediction intervals plotted on the same graph using different line types to distinguish them. The results are almost always unreadable. A better way to do this is to use semi-transparent shaded regions. Here is an example showing two sets of forecasts for the Nile River flow. library(forecast) f1 = forecast(auto.arima(Nile, lambda=0), h=20, level=95) f2 = forecast(ets(Nile), h=20, level=95) plot(f1, shadecol=rgb(0,0,1,.4), flwd=1, main="Forecasts of Nile River flow", xlab="Year", ylab="Billions of cubic metres") …