The demonstration presents a wireless system to control video games with user hand movements. Muscles activity is detected by applying the Average Threshold Crossing (ATC) technique to the surface ElectroMyoGraphic (sEMG) signals acquired from two sets of electrodes on the user forearm. Three hand movements and an idle state are classified in real-time on a computer by implementing a Neural Network (NN) feeded with the acquired ATC values, with accuracies above 97 %. Recognized gestures are then mapped to keyboard inputs to control the maneuvers of a game character.