Rogo^® is a new type of mathematical puzzle, invented in 2009. Rogo is a prize-collecting subset-selection TSP on a grid. Grid squares can be blank, forbidden, or show a reward value. The object is to accumulate the biggest score using a given number of steps in a loop around the grid. This paper introduces Rogo as a discrete optimisation problem. An IP formulation is given for the problem with two alternative sets of subtour elimination constraints. Enumeration-based algorithms are also proposed based on properties of solutions and Rogo instances. Some results of computational experiments are reported.