"I've been thinking about going to a Bollinger Band"
Bollinger has a new book out. He goes into some detail about combining M and W chart patterns with Bollinger bands.
I think Bollinger is a little too far out. He now has something he trademarked as Bollinger bars - which are just candles with fat stems.
I don't like Bollinger bands very much. They make no sense to me. I understand the concept of standard deviation and average, but when applied to a time sequence, these parameters lose their meaning. For example, I think BB would be useful if 95% of the data points (+/- 2 standard deviations) were contained between the Bollinger bands. But the way it works is that 95% of the past n data points are contained within the Bollinger band level at the leading point. So you never know for sure what percentage of points fall within the bands, and during a surge, a lot of the points fall outside the bands. I tried using Bollinger's concept of the standard deviation from the moving average to compute stop levels, but during price volatility, the stops become HUGE! and thus don't provide much protection.
However, I would like to explore M and W patterns and try to find some way to implement them into a trading system. But it is complex and will take awhile.
"Perhaps Wednesday was just one of those days."
Actually, if I had used a different set of parameters for a bullish break-out everything would have been fine. During a bear market, it should be easier to trigger a sell signal and it should be harder to trigger a buy signal. However, because I used a single set of parameters for buy and sell signals, my system triggered a buy signal too early. You referred to such a characteristic in your other post, about 1982- being bullish, just as 1998- has been bearish. When I reoptimized seperate parameters for buy and sell signals over the 1998-present data interval, a sell signal is now more easily generated than a buy signal.
My main reason for wanting to go back to old data was to try to understand the system and how the parameters would change under different market conditions, but I think I understand it a little better now. |