Bulkowski's Consecutive Closes Leading to the Breakout

Written and copyright © 2008-2011 by Thomas N. Bulkowski. All rights reserved.

Consecutive Closes Leading to the Breakout: Summary

If price closes consecutively higher/lower for more than 3 days leading to the breakout, is a throwback/pullback (respectively) more likely? No. In fact, consecutively higher/lower closes leading to the breakout are a good indicator that a retrace will not occur.

Does performance suffer? No. Post breakout performance is better if price closes higher/lower consecutively more than 3 days leading to the breakout, especially if the breakout is downward (that is, expect a larger decline).

Do failures occur more often if price closes consecutively higher/lower more than 3 days leading to the breakout? No. Failures occur half as often for upward breakouts, but only see a slight improvement after downward breakouts.

Consecutive Closes: Method

Warning (*): Please note that this research does NOT examine the price trend leading to the breakout. It only examines consecutively higher or lower closes leading to the breakout. Thus, you can have 4 days of higher closes, each a penny above the prior day, and yet price would look like it is forming a congestion zone. In that regard, this study is flawed because I was hoping to find results of price leaving a congestion area versus in the middle of a price trend leading to the breakout. What I should have looked for is price overlap leading to the breakout. Trends have little overlap but congestion areas have a large overlap.

Keep in mind as you read this study that I only found consecutively higher or lower closes, not straight-line price trends versus congestion areas. The results are still valid, but they only apply to consecutively higher or lower closes leading to the breakout. In the below text, I have marked the word trend with an asterisk to remind you of this warning. When I write trend, I really mean consecutively higher or lower closes.

I looked at 25 chart pattern types: Big M, Big W, double tops and bottoms (all types), head-and-shoulders (all types), the three triangles: ascending, descending, and symmetrical, rising and falling wedges, rectangle tops and bottoms, rounding tops and bottom, and triple tops and bottoms. I used 1,290 stocks from 7/1991 to 3/2005, but few stocks covered the entire period. I did not separate the analysis into bull or bear markets because I wanted high samples counts. The research found 12,900 samples (chart patterns).

To find the trend*, I counted the number of days that had consecutively higher or lower closes leading to, but not including, the breakout day. Any unchanged closing price was counted as part of the trend. For example, the above chart shows price making four higher closes leading to the breakout.

Consecutive Closes: Throwback or Pullback Rates

If price closes upward for more than 3 days (consecutively) leading to the breakout, does that cause a higher likelihood of a throwback? No. Throwbacks occur less often.

I found the median number of consecutive higher/lower closes leading to, but not including, the breakout was 3 days for both upward and downward breakouts. Price will throw back 70% of the time when the number of consecutively higher/lower closes is less than or equal to the median length leading to the breakout. If the number of consecutively higher/lower closes is more than 3 days (the median), then there is a 30% chance of a throwback. The analysis used 4,005 samples.

If price shows more than 3 consecutively lower closes leading to the downward breakout, will a pullback occur more often? No. Pullbacks occur less often.

For downward breakouts, the results were similar to upward breakouts. Pullbacks with consecutively lower closes less than or equal to the median 3 days had a 67% chance of occurring, but when then number of consecutively lower closes was greater than the median 3 days, the pullback rate was just 33%. The analysis used 3,318 samples.

Consecutive Closes: Performance

If price has consecutively higher/lower closes of more than 3 days leading to the breakout, will performance suffer post breakout? No. Performance improves. Upward breakouts had rises averaging 33.19% when price closed higher 3 days or less leading to the breakout. This compares to a rise of 34.53% for those chart patterns in which price closed higher more than 3 days leading to the breakout.

Similarly, downward breakouts showed declines average 18.56% versus 22.74% for consecutively lower closes less than or equal to 3 days and greater than 3 days, respectively.

In short, if price shows consecutive higher/lower closes longer than 3 days leading to the breakout, expect better performance, especially if the breakout is downward.

Consecutive Closes: Failure Rates

Do failures occur more often when price closes consecutively higher/lower more than 3 days leading to the breakout? No. They occur less often.

A failure occurs when price moves less than 5% from the breakout price and then reverses trend (moving more than 20% in the new direction). For upward breakouts in which price closed consecutively higher 3 days or less leading to the breakout, 260 chart patterns had price failing to rise at least 5%. For price closing consecutively higher for more than 3 days, the failure rate was 139 patterns, or about half.

For downward breakouts, the failure rate was 253 for price closing consecutively lower less than or equal to 3 days and 215 failed when price closed lower consecutively more than 3 days leading to the breakout.

-- Thomas Bulkowski

Copyright © 2008-2011 by Thomas N. Bulkowski. All rights reserved. The ranger isn't gonna like it, Yogi...