He needs to draw a circle and position 17 markers (the hexadecimal characters 0-9; A-F plus a question mark) around the circle. He could likely look up or perhaps even calculate that the radius of the circle inscribed inside the heptadecagon as 1/2 * cot( π/17) and the radius of the circle circumscribed around the heptadecagon as 1/2 * csc (π/17). Those evaluate to 2.674764 and 2.721096 times the length of a unit side. So the issue for Mark is how big (how large a radius) a circle does he need or can he fit? If he is free to choose the radius he might select one full step as the length of a side of the polygon. He has no way of knowing how big a circle JPL can “draw” (or that they even have a spare Pathfinder in Pasadena), and it probably does not matter much to him.
He might be better off deciding the polygon side length is something like a meter and using a piece of broken antenna or some such. Then he could measure out a piece of string 2.7 times as long and draw a circle. What he seems to have done is measure a piece of string something like just short of 5.4 paces and then “drawn” two concentric circles with his boots. The actual circumference of the heptadecagon would presumably be between the two shoes. Then he paced off the sides (two paces per side) and pounded in stakes with characters. He seems to be sitting outside the circle when awaiting the message, so we’d need to know if he can see accurately enough across the circle to tell where the camera is pointing. It seems like he’d be better off sitting right next to the camera.