motivated by email from Lynn K.
Here's an interesting notion. You take a gander at stock prices over, say, the last 10 years.
You notice that they went from P_{0} (10 years ago) to P_{1} (today).
If the Compound Annual Growth Rate is CAGR, then that means:
CAGR = (P_{1} / P_{0})^{1/10} 1 
>Huh?
That's because: P_{0} (1 + CAGR)^{10} = P_{1}.
But suppose you'd like to see how that CAGR has changed over the years.
In fact, let's look at 3year periods, where the stock price went from P_{0} to P_{1} over those 3 years.
Then the CAGR would be ...
>Don't tell me! It'd be (P_{1} / P_{0})^{1/3} 1
You got it.
Here's a spreadsheet where you can:
 Pick one of a gaggle of stocks
 Download 10 years worth of data
 Stare in awe at the variation in the CAGR (calculated over a moving 3year period).
Click on the picture to download the spreadsheet.
In the example ...
>Wait! You've got Volatility?
Yes. It's the Standard Deviation based upon 3 years worth of monthly returns, multiplied by SQRT(12) to annualize. (See square root stuff.)
In the example, IBM has been downloaded and, over the entire 10year period, the CAGR is 2.7%.
However, when calculated over 3 years, it's varied from 14.1% (over 3 years ending in Aug/03) to +21.3% (over 3 years ending in May/08).
Here are a few:
>Did you notice that Volatility seems to go in a direction opposite to the CAGR?
Fascinating, eh?
Here are a few stocks with their 3year correlations:
^DJI 80%  ^IXIC 79%  AA 77%  AIG 83%  AXP 60%  BA 79%  C 89%  CAT 57%  DD 65%  DIS 62%  F 27%  GE 85%  HD 37%  HON 86%  HPQ 85%  IBM 68%  INTC 44%  JNJ 5%  JPM 73%  KO 71%  MCD 78%  MMM 32%  MO 3%  MRK 55%  MSFT 75%  PFE 12%  PG 45%  T 93%  UTX 49%  VZ 80%  WMT 16%  XOM 83% 
