Bulkowski's Candles at Peaks and Valleys

Written by and copyright © 2005-2019 by Thomas N. Bulkowski. All rights reserved. Disclaimer: You alone are responsible for your investment decisions. See Privacy/Disclaimer for more information. Some pattern names are the registered trademarks of their respective owners.

My book, Encyclopedia of Candlestick Charts, pictured on the left, discusses candlesticks including performance statistics.

If you click on this link and then buy the book (or anything) at Amazon.com, the referral will help support this site. Thanks. -- Tom Bulkowski

$ $ $

Unusually tall price bars appear at the top of minor highs and at the bottom of minor lows. Additional research suggests that when a tall candle appears in a trend, between 67% and 72% of the time a reversal occurs within a day.

Tall Candle Methodology

I used 466 stocks of daily price data covering dates starting from November 18, 1999 to February 21, 2007, encompassing both bull and bear markets. I found all peaks at least 5 days apart and all valleys at least 5 days apart. Then I compared the height of the price bar at the peak/valley with the prior 5-day average height.

In the figure to the right, the arrows points to minor highs or minor lows that have a price bar (candle) taller than the preceding 5-day average.

Notice how price often turns within a day of the tall candle.

More

Tall Candle Results

Bull market, minor highs, candles qualifying: 33,188.

• A tall candle (including shadows) appears at the peak 67% of the time and it is 72% taller than the 5-day average.
• A tall body appears 52% of the time and it is 150% taller than the 5-day average body height.
• The candle is white 51% of the time, black 47% and no color 2% of the time.

Bear market, minor highs, candles qualifying: 17,528.

• A tall candle (including shadows) appears at the peak 66% of the time and it is 78% taller than the 5-day average.
• A tall body appears 51% of the time and it is 151% taller than the 5-day average body height.
• The candle is white 58% of the time, black 39% and no color 2% of the time.

Bull market, minor lows, candles qualifying: 33,070.

• A tall candle (including shadows) appears at a bottom 68% of the time and it is 71% taller than the 5-day average.
• A tall body appears 50% of the time and it is 144% taller than the 5-day average body height.
• The candle is white 41% of the time, black 58% and no color 2% of the time.

Bear market, minor lows, candles qualifying: 17,847.

• A tall candle (including shadows) appears at a bottom 71% of the time and it is 88% taller than the 5-day average.
• A tall body appears 54% of the time and it is 161% taller than the 5-day average body height.
• The candle is white 40% of the time, black 57% and no color 3% of the time.

More

Methodology for Tall Candle Reversals

In a manner similar to the above methodology, I found all candlesticks that were taller than the 146% median for tall candles (the 146% value comes from the above testing. I found the height for candles taller than the 5-day average and then used the median height of those tall candles). To check for price trending, I included a simple test to compare today's high price with the high prices of 2 and 3 days ago. This helped to assure the tall candle was at or near a minor high. The prior day was not checked because I wanted to include it in the one-day window (that is, the day before the tall candle could be the peak). In a similar manner, I checked the low price of days 2 and 3 before the tall candle to make sure the tall candle had the lowest low and skipped the prior day.

I compared the tall candle signal with a minor high separated from other highs by at least 2 days or a minor low separated from other lows by at least 2 days.

The test used data from 466 stocks beginning from approximately 10/20/1999 to February 22, 2007, encompassing both bull and bear markets. I used a 22 day average of the daily high-low range to determine whether a candle was at least as tall as 146% of the average. The 22 day (1 month of trading) average results in the best performance of 3-5, 10, and 22 day periods. If so, then I looked to see if a minor low or minor high was nearby (no more than 1 day away, either before, during or after the tall candle).

Tall Candle Reversal Results

Minor Highs: 58,676 samples

• A tall candle marked a minor high peak 39% of the time.
• 17% of the time, the next day marked the minor high peak.
• 11% of the time, the prior day marked the minor high peak.
• Together, a tall candle marked a minor high 67% of the time, plus or minus 1 day.

Minor Lows: 53,391 samples

• A tall candle marked a minor low valley 42% of the time.
• 17% of the time, the next day market the minor low valley.
• 12% of the time, the prior day market the minor low valley
• Together, a tall candle marked a minor low 72% of the time, plus or minus 1 day.

More

Tall Candle Trading

If you see an unusually tall candle...

1. The high of the tall candle must be above the high's of 2 and 3 days ago (for potential peaks) or below the lows of 2 and 3 days ago (for potential valleys).
2. Compute the average height of the preceding 22 trading days (using the difference between the high and low for each day and then averaging them). Do not include the tall candle in the computation.
3. Multiply the average height by 146% and compare it to the height of the tall candle. Tall candles should be above the computed result.
4. If the candle is taller than the average then expect price to peak or valley within a day between 67% (peaks) and 72% (valleys) of the time.

For example, the above figure shows all of the signals indicating tall candles at minor highs (down arrows) and minor lows (up arrows). Red candles are the mistakes. Remember that the signals are valid from one day before to one day after the tall candle.

-- Thomas Bulkowski

More

Written by and copyright © 2005-2019 by Thomas N. Bulkowski. All rights reserved. Disclaimer: You alone are responsible for your investment decisions. See Privacy/Disclaimer for more information. Some pattern names are the registered trademarks of their respective owners. Good judgment comes from experience, and experience comes from bad judgment.