If you're going to "unveil a robot", then a better looking Scissors would be called for here. The algorithm is also likely foolable by throwing what appears to

If you're going to "unveil a robot", then a better looking Scissors would be called for here. The algorithm is also likely foolable by throwing what appears to be one symbol (e.g. rock), but then morphs into another symbol (two fingers go out late in the symbol formation, in order to present scissors upon completion of the movement). A comparison of the reaction time of this robot vs. say... Google's automonous driving car, or vs. an automatic stock trading machine, might present the robot as interesting. Multiple cameras and a wide variety of acceptable symbol geometries on the part of the human player (rock rotated 90 degrees, the use of two hands, etc) would add realistic complexity... and would require more complex recognition algorithms, likely slowing down the reaction time of the robot.