1) Download historical data for the S&P 500 index (or use VFINX as a proxy).
Choose a random 365-day period.  Graph the closing prices on these days, scaled
by the initial closing price (so that every graph starts at 1.00).

2) Generate Lognormal differentials with mean 0.00001 and standard deviation[*]
0.005. Graph the corresponding stock chart.

3) Do it again, with mean -0.00001 and the same standard deviation.  Graph
the corresponding stock chart.

Now, here's the experiment part.  See if your neighbor can figure out which
one is the real chart.