Bulkowski on the 2B Pattern (2024)

HomeBlogFAQs

Busted Index Pic Index Event Indicator Rank Small Ugly

Candle Index Candle Pic Index Elliott Wave Volume

Books Class

Quizzes Setups Studies Traps Tutorials

About Forecast Fundamentals Glossary Scoring Software

What's New

2B: Summary

The 2B pattern works as Sperandeo says. For peaks, if a first peak is slightly below the second one, expect price to drop further.

For valleys, if the first valley is slightly above the second, expect a larger rise.

I show the two variations in the figure.

Enter Trader Vic and 2B Pattern

Victor Sperandeo in his book,Trader Vic--Methods of a Wall Street Masterdescribes the 2B pattern this way.

In an uptrend, if a higher high is made but fails to carry through, and then prices drop below the previous high, then the trend is apt to reverse. The converse is true for down trends.This observation applies in any of the three trends; short-term, intermediate-term, or long-term.
A 2B on a minor high or low will usually occur within one day or less of the timethe high or low is made. For 2B's on intermediate highs or lows preceding a correction, the new high or low point will usually break within three to five days. At major market turningpoints, long-term 2B's, the new high or low will usually break within seven to ten days. In the stock market, after the new high is made, the failure to carry forward usuallyoccurs on low to normal volume, and the confirmation of a reversal occurs on higher volume.

I prefer to think of a 2B as one in which price begins to form a double top. It may not confirm as a double top (meaning that price may not close below the price ofthe valley between the two peaks) but price exceeds the level of the first top and then reverses.

A 2B Pattern Example

Let's look at a chart example. I show Omnicom Group (OMC) in late 2016 on the daily scale. Price rises in a strong upward trend (red line A) to peak B.

The stock retraces a good portion of the move up (that is, it drops and forms the valley between peaks BC).

2B Top: Research

I programmed my computer to find all peaks in a 10-year period from January 1, 2010 to January 1, 2020. I chose that period because it was after the March 9, 2009 bear market and beforethe COVID-19 pandemic affected the stock market. I used 468 stocks and logged 41,702 samples from stocks priced above $5 a share at the top of the first peak.

Here's what I found.

Does the 2B actually work? Yes. When price rises above the first peak and reverses, the stock will drop more than if the stock peaks below the first peak. Here's the numbers.

Description	First Peak Below Second					Equal	First Peak Above Second
Peak-to-peak difference:	-4%	-3%	-2%	-1%	0%	0%	1%	2%	3%	4%	>4%
Percentage loss:	16%	14%	12%	10%	7%	6%	6%	6%	6%	6%	7%
Samples:	199	322	685	1,872	6,456	74	10,982	8,082	5,676	4,144	3,210

For example, when the first peak was below the second by 4% or more, the 199 samples saw price drop an average of 16% (measured from the second peak's high to the ultimate low). At the far right of the table, when the first peak was more than 4% above the second one, the 3,210 samples saw price drop an average of 7%. I limited the peak-to-peak price difference to 5% (below or above the first peak). The median drop for all samples was 7%.

Peaks closer together saw price drop farther. The median time between peaks was 10 days. Sorting the results into less than and greater than the median showed that peaks closer together lost more: 7% versus 6%, respectively.
I checked all peaks and found that price dropped a median (and average) of 10%. I consider this the "benchmark" number. It's the amount I'd expect to see regardless if a 2B formed at a peak or not. I found the benchmark number for each stock before I made a comparison for that stock. Thebenchmark number ranged from 5% to 22%, depending on the stock. That means price could drop between 5% and 22% after the second peak (measure from the high at the second peak to the ultimate low) andstill qualify as a normal decline, depending on the stock. I found that the 2B pattern beat the benchmark group (by seeing price drop farther) only 34% of the time. I find that result puzzling.
Splitting the results into two, those with the first peak below the second showed 2B patterns beat the benchmark group 42% of the time. That compares to a 32% win over the benchmark group when the first peak was above the second one.

To put these results another way, seeing a second peak form within 5% above the first peak allows you to exit the stock and avoid an above average decline 42% of the time. The other 58% of the time you'd see an average decline of 10%.

The 2B pattern, where the second peak is slightly above the first peak before dropping, results in a larger loss than if the second peak were at or below the first peak.This is just as Sperandeo said in his book.

2B Bottom: Research

In a similar manner to 2B tops, I programmed my computer to find 2B bottoms. That's when price forms two bottoms with the second bottom near the first bottom in price.

I looked for a second bottom that was within 5% of the first bottom and up to a year in width. I found 14,890 examples in 468 stocks using data from January 1, 2010 to January 1, 2020.The time span was to avoid a bear market and Covid-19 reaction in the stock market.

Description	First Valley Above Second					Equal	First Valley Below Second
Valley-to-valley difference:	-4%	-3%	-2%	-1%	0%	0%	1%	2%	3%	4%	>4%
Percentage gain:	43%	40%	40%	37%	34%	21%	12%	32%	30%	32%	30%
Samples:	175	256	475	1,018	2,275	10	6	3,279	2,869	2,439	2,088

Let's discuss a few examples from the table. I found that when the first valley was above the second by 4% (shown as -4%, meaning price dropped from valley to valley), the percentage gain from 175 samples was 43%. Let's lookat the right side of the table. When the first valley was below the second by 4%, the gain from the second bottom's low to the ultimate high was 32% from 2,439 examples.

In short, the 2B pattern also works for bottoms. When the second valley is slightly below the first (first valley above the second), expect a bigger gain.

Here are additional findings.

Performance improves if the bottom to bottom time is less than 21 days. The 21 number is the median time between bottoms for all 2B patterns. I found that the average gain was 33% for those less than or equal to 21 days and 32% for those with bottoms separated by more than 21 days.
A first bottom above the second performs more strongly when sorted by the 21-day median. Splitting the patterns into those shorter or longer than 21-days wide andwhether the first bottom was above or below the second, I found substantial performance differences. For patterns less than or equal to 21 days wide (bottom to bottom) and the first valleywas above the second, performance was 36% compared to 31% for those less than or equal to the 21-day median but had a second valley below the first. For those with bottoms more thanthe 21-day median, performance was 37% versus 31% for the first valley above/below the second, respectively.
I found that the normal rise after a bottom averaged 34% but this varied from stock to stock with a range of 13% to 183%. A count of the times a 2B pattern (first valley above the second)beat the 34% average for all bottoms was just 33% of the time. This compares to a 26% win rate when the first valley was below the second. This means that if you trade a 2B bottom, you can expect an above average gain 33% of the time and an average gain the other 67% of the time.

I ran the same simulation using wider-spaced bottoms with a median of 59 days apart. The trend of the results was similar although the values were different. The one exception wasto the comparison of less or more than the 59-day median versus the first bottom above/below second one. In this case, I found that when the first valley was above the second performancewas worse than those with the first valley below the second, with gains averaging 51% to 54%, respectively. This is a flip from narrow (21-day median width) patterns.

-- Thomas Bulkowski