Bulkowski’s Falling Window

Written and copyright © 2008-2009 by Thomas N. Bulkowski. All rights reserved.

In my book, Encyclopedia of Candlestick Charts, pictured on the right, I explore the entire range of candlestick patterns from abandoned babies to windows (not exactly A to Z, but you get the idea), in both bull and bear markets, using almost 5 million candle lines in the tests.

The book takes an in-depth look at 103 candlestick patterns and reports on behavior and rank (3 types: reversal rate, frequency, and overall performance), identification guidelines, performance statistics (tables of general statistics, height, and volume), trading tactics (tables of statistics on reversal rates and performance indicators), and wraps each chapter with a sample trade. I share a sliver of that information below. If you like what you read here, then you will love the book. Help support this website and buy a copy by clicking on the above link.

The falling window is a fancy name for a price gap in a downward price trend. It occurs when yesterday’s low is above today’s high, leaving a hole on the daily price chart. The pattern appears in a falling price trend, and it acts as a bearish continuation pattern. Falling windows occur often, so you will find them on the charts, but the longer the time scale, the more difficult it is to find one. Once the trend is underway, it tends to remain moving as the overall performance rank suggests.

Important Results

Theoretical performance: Bearish continuation Tested performance: Bearish continuation 67% of the time Stopped in gap: 25% Frequency rank: 23 Overall performance rank: 7 Average time to gap closed: 55 days Median time to gap closed: 9 days

Falling Window

Discussion

As I mentioned in the introduction, the falling window is a space left on the price chart. On the daily chart, a dividend can cause a gap as well as a surprising earnings announcement or many other types of events. The falling window acts as a bearish continuation pattern 67% of the time, which is respectable. The overall performance rank is 7, but that really measures the price trend surrounding the falling window and not the window itself.

One of the more interesting statistics from Important Results is the stopped in gap number. For a falling window, this is the percentage of time that a minor high appeared within the gap before price closed the gap. In other words, a gap showed overhead resistance 25% of the time.

The average time to close the gap is 55 days, but the median is 9 days. Closing the gap means price retraces far enough to cover the gap. If price gaps from $1.50 to $1, then price would have to rise back to $1.50 to close the gap. The large difference between the two numbers, 55 and 9, is because the average has some gaps which take a long time to close, pulling the average upward. The median just splits the list in two and reports on the middle number.

Identification Guidelines

Example

The chart of 3M shows many different windows, some falling and some rising. A falling window will appear in a downward price trend, such as that shown at point A. The day after the white candle, price gaps open lower and struggles to close the gap throughout the day, but cannot do it. A hole remains on the chart.

The other points, B, C, and D, are all gaps called rising windows. Those have a high price on one day that remains below the low of the next day, leaving a hole on the chart.

One of the secrets to rising and falling windows is to determine the gap type. Rising window B, for example, is a breakaway gap because it breaks away from the small congestion area. C is an exhaustion gap because the price trend ends soon after. Point D is an area or common gap because price closes that gap shortly after it appears.

What type of gap is falling window A? Since price is trending lower, it is probably an exhaustion gap and not an area gap.

-- Thomas Bulkowski