As of 09/06/2024
Indus: 40,345 410.34 1.0%
Trans: 15,427 190.44 1.2%
Utils: 1,029 7.87 0.8%
Nasdaq: 16,691 436.83 2.6%
S&P 500: 5,408 94.99 1.7%

YTD
+7.0%
3.0%
+16.7%
+11.2%
+13.4%

42,000 or 39,500 by 09/15/2024
16,200 or 15,300 by 09/15/2024
1,100 or 1,000 by 09/15/2024
18,400 or 16,650 by 09/15/2024
5,750 or 5,375 by 09/15/2024

As of 09/06/2024
Indus: 40,345 410.34 1.0%
Trans: 15,427 190.44 1.2%
Utils: 1,029 7.87 0.8%
Nasdaq: 16,691 436.83 2.6%
S&P 500: 5,408 94.99 1.7%

YTD
+7.0%
3.0%
+16.7%
+11.2%
+13.4%
 
42,000 or 39,500 by 09/15/2024
16,200 or 15,300 by 09/15/2024
1,100 or 1,000 by 09/15/2024
18,400 or 16,650 by 09/15/2024
5,750 or 5,375 by 09/15/2024
 
Initial release: 8/9/2024.
Imagine you're following a company and its stock gaps down. Soon, a chart pattern appears. Is there a higher likelihood of the pattern failing because of the gap? What if the move down is a tall price bar? Do tall gaps and tall price bars affect future performance?
To answer these questions I conducted research. What I found is that performance was affected, but you probably won't notice it in your trading.
Here's the results.
Let's show an example of this situation. The adjacent picture shows a chart pattern called an Adam & Eve double bottom (marked as AE on the chart). It becomes a real chart pattern after the stock closes above the top of the chart pattern.
In this study, the gap must appear within 21 price bars (about a month) before the end of the chart pattern (before bottom E), which it does. I did this because I wanted to understand if nearby tall bars or gaps influenced trade success.
In this example, the breakout is upward (in all tests, I only looked at upward breakouts). The top of the pattern is at 46.73, the bottom is at 39.56 for a height of 7.17. Added to the top of the pattern gives a target, shown, of 53.90. The stock exceeded the target, so this was classified as a winning trade. I show a stop loss as a penny below the bottom of the chart pattern.
A losing trade would see the stock price drop far enough to hit the stoploss order.
The data I collected started from October 1990 to August 1, 2024. I found 21,107 samples from bull markets.
The median gap height was 2% (with a 1% minimum). The median distance from the gap to the end of the chart patterns was 18 calendar (not trading) days (the maximum allowed was about a month).
If you look at chart patterns that did NOT have a tall gap within a month of the pattern's end, the failure rate was 30%. That means almost a third did not see price climb to the target before being stopped out for a loss.
When a nearby bullish gap (when price gaps higher) appears, the failure rate drops from 30% to 29%. That's not a difference you'd notice, but it says that tall bullish gaps improve performance (slightly).
The chart shows failures versus time. As we increase the distance from a tall bullish gap to the end of the chart pattern, we see failure rates drop (for a time).
The failure rate remains steady at about 29% for most of the chart until day 24 where failures drop (and so do samples, 629 at day 24 to a low of 107 at day 29).
What I can't explain about this chart is when a bullish gap appears close to a month ago, failure rates drop. One explanation is that the gap fills (price retraces far enough to cover the hole on the chart) and then rebounds. If price rises far enough to reach the target, you'll lower the failure rate.
I used the same method to study bearish gaps, where price gaps lower.
The chart shows an example. This is an Adam & Adam double bottom (marked as AA on the chart). The top of the pattern is at 7.17, the bottom is at 6.47, for a height of $0.70. Added to the top gives a target of 7.87. The stock reached the target quickly and closed the gap.
This looks like a continuation gap where the gap appears midway along the trend and it takes longer to close, averaging 45 days.
The stock, after the double bottom, climbed far enough to reach the target, so it was a winning trade.
I found 21,133 samples where price gapped lower according to the below Methodology. The median distance from the gap to the end of the chart pattern was 20 days with a maximum of a month. The median gap height was 2%, with a 1% minimum.
As the chart shows, bearish gaps close to the end of the pattern (near when you'd buy, hoping for a rise) has a higher failure rate than those trades with gaps farther apart.
For example, the first and highest point on the chart says that for bearish gaps that appear the day before the chart pattern ends, bullish trades (101 of them) failed 41% of the time compared to a failure rate of 30% for those gaps farther away (2,525 trades).
As the distance from the gap to the chart pattern increases, failure rates decline and hold steady near 30%. For those chart patterns without a gap preceding them, the failure rate is 30%.
Median  Failure Rate 
<=20 days from gap to chart pattern end  31% 
> 20 days  29% 
Short gaps (<= 2% high)  30% 
Tall gaps (>2%)  31% 
No gaps  30% 
When you compare failure rates with the median distance (20 days), we see the results shown in Table 1.
Nearby bearish gaps have a higher failure rate than those farther away. Tall gaps are slightly more bearish than short ones.
The chart shows a tall bullish price bar within an Eve & Eve double bottom chart pattern.
For this test, the tall price bar must be within a month from the end of the chart pattern. Having it appear within the pattern may look odd, but it's fine.
The height of the double bottom is $23.45 for the top, $17.47 for the bottom, giving a height of $5.98. Adding the height to the top of the chart pattern gives a target of 29.43, as shown.
The stock met and exceeded the target without sweating it.
I collected 21,139 samples. The median failure rate of all chart patterns not preceded by a tall bar is 30%. When a tall bar appears, regardless of its height (up to 4 times the average size over 2 months) and regardless of its distance (up to a month away from the pattern's end), the failure rate remains about 31%.
I thought failures would drop after a bullish bar, but they don't.
A tall bullish price bar has no influence on future performance.
This chart shows an Adam & Adam double bottom chart pattern (AA) with a tall price bar that's bearish (meaning it closed lower than the prior day).
Does the tall and bearish price bar help predict better or worse future performance?
To measure that, I conducted similar tests as already described.
In this example, the top of the chart pattern is at 38.20, the bottom is at 33.84 for a height of 4.38. Add the height to the top of the chart pattern gives a target of 42.56, as shown on the chart. The stock reached the target.
I used a stop loss of a penny below the bottom of the chart pattern to close out losing trades.
In this test, I found 21,118 samples (trades) where the price bar within a month of the end of the chart pattern was at least twice the height of the 2month average.
The median of the bars I logged was 6% tall and 19 days away from the end of the chart pattern.
Tall price bars closer than the median 19 days failed 30% of the time. Those farther away failed slightly less, 29% of the time. Nearby tall bearish bars are more bearish than those farther away.
Chart patterns without tall price bars failed 30% of the time. When multiple tall price bars appeared (within a month of pattern's end), the failure rate climbed to 31%.
Table 2 shows the different failure rates as tall price bars grow in size. See the Methodology for more details on how I found the results.
Here's how to read the table. Let's discuss the 3.00 row. I sorted trades into two buckets, one bucket held all "tall" price bars that were equal to or shorter than 3 times their average height, and the other held the tall ones.
I found that short bars led to a failure rate of 30% compared to a failure rate of 29% for their taller counterparts.
If you look down the rows, you'll see that tall bars have a lower failure rate than shorter ones. Keep in mind that as the height of the price bar climbs, sample trades drop (down to 137 for the last line).
Multiplier  <= Size  > Size 
2.25  31%  29% 
2.50  31%  28% 
2.75  30%  29% 
3.00  30%  29% 
3.25  30%  27% 
3.50  30%  27% 
3.75  30%  27% 
4.00  30%  25% 
Do the results make sense? No. Why would a tall, bearish price bar lead to a lower failure rate than a short bar? The only explanation that I can think of is that the tall price bar forms a chart pattern called a spike or tail.
In those patterns, the stock opens high, makes a new low but closes the day near where it began (high), leaving a tall (downward) price spike on the chart. It's as if the bears force down price and the bulls send it back up. It's a bullish pattern, which is more reliable than when price makes a tall upward spike.
Overview: In the test, I found unusually tall gaps or price bars before the end of a chart pattern and measured the rise to a price target or a drop to a stop loss order.
I used a similar methodology to study both tall price bars and tall gaps. In short, I used the height of a chart pattern added to the top of the pattern as the target. If price climbed to the target, then the trade was a winner. If price first touched a stop loss order (a penny below the bottom of the chart pattern), then the trade was a failure.
Below are the details.
 Thomas Bulkowski
Support this site! Clicking any of the books (below) takes you to
Amazon.com If you buy ANYTHING while there, they pay for the referral.
Legal notice for paid links: "As an Amazon Associate I earn from qualifying purchases."
My Stock Market Books

My Novels
