By inscribing a circle in a square and then a square in a circle find an approximation to pi.
By using a hexagon, can you improve on the approximation? How much better an approximation is it?

Archimedes used this idea first with a hexagon, then a dodecagon (12 sides) and so on up to a 96 sided polygon to calculate pi and was able to establish that
3 10 71 <π<3 1 7 .

What are the strengths and limitations of this method?