I finally got some time this weekend to look at using $NYSI 5,3 stochastic crossovers to time entries and exits on the S&P 500 weekly. Over the past two years, these crossovers appeared to give fantastic signals, and even when considering how you would actually have to execute trades, the results are pretty good (65% winning trades and a 22.2% CAGR). The bad news is that the equity curve over those two years would have made it tough to keep on the system in that you would have been negative for the first nine months. More of a problem is that the same methodology backtested further breaks down pretty badly, with lots of spurious buy and sell signals during the extended 2006 rally.

The method I tested was to buy / sell at the closing price of the S&P 500 on any Friday where the $NYSI 5,3 stochastic crossed up / down. I worked my way week by week through the weekly $NYSI chart taking a screen shot for every signal, noting the closing S&P 500 price. I acted on every reversal crossover—including the nasty whipsaw in late Nov 2008 / early Dec 2008.

The shots below are 1) Matt’s signal chart 2) my revised signal chart 3) the list of trades 4) the equity curve and 5) a picture of the first signal on 10-19-2007 that also shows the spurious signals given during the long 2006 rally.

While I’d be awfully happy with a compound annual growth rate of 22.2%, I don’t think you’d get those kind of numbers if you look back further into time (I plan to try this eventually).

-pd