We study the control of large scale energy storage operating in a market. Re-optimization of deterministic models is a common pragmatic approach when prices are stochastic. We apply Lagrangian theory to develop such a model and to establish decision and forecast horizons when storage trading affects these prices, an important aspect of some energy markets. The determination of these horizons also provides a simple and efficient algorithm for the determination of the optimal control. The forecast horizons vary between one and fifteen days in realistic electricity storage examples. These examples suggest that modelling price impact is important.