What R package do you use when you want to produce forecasts for a time series? The answer would probably be forecast. While this is a very good package, that contains a lot of useful functions, some of them are rather restrictive. smooth package for R contains several more flexible functions for forecasting purposes. The list of functions includes:

- es() – the function that implements ETS (exponential smoothing in Error-Trend-Seasonal taxonomy) model. The advantages of the function in comparison with ets() function from forecast package are discussed here;
- ssarima() – state-space ARIMA model, the function that implements versatile and flexible ARIMA that allows estimating simple and seasonal ARIMA using advanced methods;
- msarima() – Multiple Seasonal ARIMA, that allows constructing models with several seasonalities (e.g. for high frequency data) in a reasonable time;
- ces() – Complex Exponential Smoothing, model that side steps the ETS taxonomy and allows producing non-linear forecasting trajectories;
- gum() – Generalised Univariate Model – model that underlies them all;
- sma() and cma() – Simple Moving Average and Centered Moving Average respectively
- simulate() – function that allow simulating data from different models

Some information about the package can be found on its page on github.