motivated by a discussion on diehards.org
The big question (often asked ... and answered) is: Should one rebalance a portfolio to maintain some asset allocation which ...
>Some allocation that you feel is best, right?
Uh ... yes. Maybe it's balanced between stocks and bonds and domestic and foreign and small caps and large caps and ...
>Yeah, okay. So what?
So I thought it'd be neat if, instead of arguments based upon financial logic or mathematical wizardry, we pick annual returns from historical data and see what'd happen.
>Are we talkin' Monte Carlo?
Why not?
 You pick a 4asset portfolio and the allocations and the spreadsheet has annual returns for each asset, from 1928 to year 2000.
 If you press F9, a random year is selected and the returns for that year applied to the portfolios.
 This is repeated 25 times and you get to see a pair of 25year portfolios: reBalanced or NOT reBalanced.
 When you get tired of pressing F9, you click a button to Do the Monte and the spreadsheet presses F9 umpteen times for you ... a la Monte Carlo.
Then you get to see how often reBalancing wins ... or not.
 If you'd like to see a specific 25year time period (with returns occurring in the "actual" historical order), there's a button for that, too.
The random selection of returns gets turned off ... but you can turn the randomizing back on.
>Why just 1928 to 2000?
Them's the numbers I happen to have.
>And do they include foreign assets, like maybe ...?
Uh ... no, but you can add your own annual returns way over in columns AD and/or AE or replace the annual returns in the spreadsheet with your own.
>So is reBalancing a good idea?
How would I know? Play with the spreadsheet ...
>I assume that everybody rebalances, right?
Actually, it's more like 82.3% and ...
>That's exact?
Of course! Would I kid you?
 
Some interesting charts:
