mechinv wrote: The Buy and Sell strategy has the following rules.
1. Buy the S&P 500 whenever it closes at a 6 month high.
2. Sell the S&P 500 whenever it closes at a 1 year low, and put the proceeds into 5 year treasuries.

This is not exactly what he wrote in the link you provided.

9. The Buy and Sell Portfolio:

Buy when DJIA/S&P 500 closes at a 6 month high. Sell DJIA/S&P 500 closes at a 1 year low. Put the proceeds into 5 year treasuries.
CAGR: 1926-2010= 12.77% ( worst drawdown=-20.30%)( worst drawdown for buy and hold=-64.21%)
CAGR: 1972-2010=13.84%( worst drawdown=6.09%)( worst drawdown for buy and hold=-37.61%)

This immediately did not pass the sniff test. Worst drawdown is 6.09%! Very doubtful. So I wondered what it would take me to plug in the data and run the test so I used my stopwatch - total time 5 minutes and 55 seconds. I'm not convinced that AI can beat out a simple setup in Excel. Not yet, anyway!

So the results? First off, you'll note that he states DJIA/S&P 500 closes. I'm not sure what he means so I tested the S&P 500 by itself, the Dow by itself, and choosing the earliest signal of the two. It may be that he means averaging the two indexes to produce the signal. Not sure.

First, we test 1972 to 2010.

S&P 500 Signals

* CAGR 10.80%
* Max DD -32.47%

Dow30 Signals

* CAGR 8.57%
* Max DD -42.63%

Combined Signals

* CAGR 9.19%
* Max DD -39.60%

Second, we test 1972 to present.

S&P 500 Signals

* CAGR 10.74%
* Max DD -32.47%

Dow30 Signals

* CAGR 8.96%
* Max DD -42.63%

Combined Signals

* CAGR 9.77%
* Max DD -39.60%

It is very simple to realize the Max DD he claims is impossible. Why? The 1987 drop! The sell is 10/19/1987 in all three cases. The buy back into the market is 6/14/1988. The Drawdown period is 8/25/1987 to 8/22/1988.

I wish he had posted his actual signals and then it would be easier to see where he failed in his setup.