I decided to adopt Schaeffer's _Activity Based Statistics_ as a resource text for labs, and just did the one called "Spinning Coins" with the kids. The counterintuitive distribution of % of heads when spinning rather than tossing was well made by the lab, but I threw in a twist which I thought you might like. A colleague told me that a 1964D penny was the closest thing to an honest "loaded" coin because of the extra recessing of Lincoln's portrait on the heads side, so we found a couple and spun those for the lab. Sure enough, even though the total distribution had a mean of 39.8% heads (not a surprise; n = 25) the distribution for the '64D had a mean of _29%_! (n = 20) I thought this was pretty cool, and would like to do some tossing with the same coin to see what happens. Anybody have any experience with this?