Predictions
motivated by e-mail

Here's what we want to do:
 Download ten years worth of daily prices for some stock, say GE. Generate a distribution of daily returns, like Figure 1. Apply this distribution of returns to a fictitious stock whose current price is, say \$10.     This fictitious stock has nothing to do with GE stock.     Its evolution in time just uses the GE return distribution. Pick a number of days into the future, say 30. Select 30 returns at random, from the return distribution (as in Fig. 1). See what happens to your \$10 stock, after 30 days. Repeat steps 5 and 6 a jillion times and generate a probability distribution of prices, 30 days in the future. Figure 1
>And that's exact?
Are you kidding? To be exact you'd need this.

It's an estimate of a future stock price assuming the historical return distribution of GE stock is relevant. Now that you have an estimate ...

Yes. Patience!
Having an estimate of a future price such as Figure 2, you might use it to estimate the value of a stock option, for example. In Fig. 2, the red dot is the current price that you've entered.

>Options?
Black-Scholes says:
 Call Premium = S*NORMSDIST((LN(S/K)+(R+V^2/2)*T)/(V*SQRT(T))) - K*EXP(-R*T)*NORMSDIST((LN(S/K)+(R+V^2/2)*T)/(V*SQRT(T))-V*SQRT(T))

Figure 2
>Huh?
The various parameters involved are explained here and ...