As of 09/29/2023
Indus: 33,508 158.84 0.5%
Trans: 14,969 17.13 0.1%
Utils: 817 +1.02 +0.1%
Nasdaq: 13,219 +18.04 +0.1%
S&P 500: 4,288 11.65 0.3%

YTD
+1.1%
+11.8%
15.6%
+26.3%
+11.7%

34,500 or 33,150 by 10/15/2023
15,500 or 14,200 by 10/15/2023
880 or 800 by 10/15/2023
13,750 or 12,700 by 10/15/2023
4,425 or 4,175 by 10/15/2023

As of 09/29/2023
Indus: 33,508 158.84 0.5%
Trans: 14,969 17.13 0.1%
Utils: 817 +1.02 +0.1%
Nasdaq: 13,219 +18.04 +0.1%
S&P 500: 4,288 11.65 0.3%

YTD
+1.1%
+11.8%
15.6%
+26.3%
+11.7%
 
34,500 or 33,150 by 10/15/2023
15,500 or 14,200 by 10/15/2023
880 or 800 by 10/15/2023
13,750 or 12,700 by 10/15/2023
4,425 or 4,175 by 10/15/2023
 
Initial release 10/30/2020.
I ran a test to see which of ten types of shortterm reversal patterns resulted in the biggest average decline. The answer is what I call the Below 3 pattern. In that pattern, I looked for a high price above the following 3 days. If those three days remained below the peak of the first bar, then I entered the trade at the open the following price bar.
Here's the list of patterns tested, ranked from best to worst by the percentage profit.
Rank  Test 
1 (Best)  Below 3 
2  Openclose reversal, uptrend 
3  Inverted 3LR 
4  Pivot point reversal, uptrend 
5  Closing price reversal, uptrend 
6  Hook reversal, uptrend 
7  Below 2 
8  Key reversal, uptrend 
9  Pivot 
10 (worst)  Below 1 
Here are the rules I followed to test these patterns.
Some of these rules are made clearer below.
Consider the chart on the right.
The three red bars form an inverted 3LR pattern. That's two higher peaks followed by a drop (C) with a low below the first red price bar.
The day before the first red price bar, A, I checked the slope of the linear regression line found using the ten black price bars (starting from F) leading to and including A. In this example, the line slopes upward, confirmed with the highlow average of the first black bar (F) compared to the highlow average of the last price bar in the ten (A).
The inverted 3LR appears and makes a high at B followed by the end of the pattern at C. Price at E recovered and posted a high above the high price at B. If E remained below B until the end of data, then I didn't include the trade in the results.
I measured the largest drop between bar C and E using the low at C and D, and divided the difference by the low at C to get the percentage drop.
The following table shows how the reversals performed. Note that this is for perfect trades. It measures the drop from the opening price the day after the pattern ends to the lowest low before price begins its recovery.
As you scan down the list, notice that the percentage profit (6% to 8%) and the range between the two is small. Failures are high, too, usually in the 60s. Profit per share ranges from a low of $2.45 to a high of $4.18.
In other words, there's not a big difference in performance.
Test  Profit  Samples  % Profit  5% Failures (Rank)  Profit/share (Rank) 
Below 3  $95,885.63  31,330  8%  62% (1 best)  $3.06 (4) 
Openclose reversal, uptrend  $851.51  348  8%  71% (10 worst)  $2.45 (10 worst) 
Inverted 3LR  $11,625.22  2,783  7%  63% (2)  $4.18 (1 best) 
Pivot point reversal, uptrend  $63,188.99  20,969  7%  63% (2)  $3.01 (5) 
Closing price reversal, uptrend  $35,537.59  12,953  7%  64% (4)  $2.74 (7) 
Hook reversal, uptrend  $2,900.05  1,136  7%  65% (5)  $2.55 (8) 
Below 2  $106,943.00  37,722  7%  65% (5)  $2.84 (6) 
Key reversal, uptrend  $5,175.99  1,605  7%  66% (7)  $3.22 (3) 
Pivot  $16,056.64  4,874  7%  66% (7)  $3.29 (2) 
Below 1  $126,000.30  50,120  6%  69% (9)  $2.51 (9) 
Here are the 10 patterns I tested, in order of performance, from best to worst.
This is a pattern I use to find peaks and valleys, using various number of price bars (usually 3, 5 or 10, but I tested only 1, 2, or 3) after a peak or valley.
For the Below 3 pattern, find the tallest peak followed by three price bars which don't exceed the first bar's high price. Thus, the first price bar is the highest of the four. Enter at the open a day after the last bar.
In the figure, bar A is higher than 1, 2, and 3. I don't care about the low price or anything else except the high price of each bar.
In this test (Below 3 pattern), I found 31,330 of them. Price dropped an average of 8% which ties for first place and 62% of the patterns failed to drop more than 5%. That high failure rate is not startling for short patterns. Even so, the failure rate places the pattern first (best). The trades made an average of $3.06 a share.
With this pattern, you're entering the trade on day 5 which probably means price is already heading lower. That keeps failures low but also hurts profit per share.
For the Below 2 pattern, it's the same pattern except I'm only looking at bars 1 and 2 compared to A. Bars 1 and 2 must have a high below bar A.
For Below 1, it's the same except I'm comparing price bar A with 1. Bar 1 must have a lower high price than A.
In this pattern, price opens near the high price for the day and closes near the low for the day but above the prior close.
It's easier to explain it with formulas.
The chart on the right shows an example. The openclose reversal occurs in an uptrend. Price opens near the high (red bar), closes near the low (brown bar), and closes above the prior close (green bar).
In my tests, I found only 348 of them but they performed well in terms of percentage profit (8%, tied for first place). However, 71% failed to see price drop more than 5%. That's the worst failure rate of the tests. Each share gained an average of $2.45, which is also the worst performance.
The figure on the right shows the inverted 3LR.
I found out about the inverted 3LR pattern from an article in Traders.com (a printed magazine, not the website) and according to the author Paolo Pezzutti, it's based on the work of Michael Harris.
3LR stand for 3 lows and a reversal. In this case, however, we're looking for 2 higher highs and a reversal with a low below the first price bar in the pattern.
In my test, I found 2,783 of them which made an average of 7% with 63% failing to see price drop at least 5% (tied for second place). Each share made on average $4.18, the highest profit per share of the tests. This is a rare pattern but you enter a trade close to the high, which boosts pershare profit.
This pattern is simple enough and I show a picture of it on the right.
In an uptrend, price closes below yesterday's low. I also added a rule that today's high price must be below yesterday's high. The rule I added isn't part of the pattern, according to my web page describing the pattern.
Testing shows this to be a plentiful pattern. I found almost 21,000 of them. They made an average of 7%, failed 63% of the time (tied for second), and made $3.01 a share.
This looks like a variation of the openclose reversal, uptrend pattern.
See the chart to the right.
Look for the opening price to be near the intraday high. The close must be near the intraday low and below the prior close.
In the figure, the horizontal red bar is the prior close. The pink bar is today's open, which must be near the intraday high (within 25% of the price bar's height, below the top), and the cyan bar is today's close. It must be below the red bar (the prior day's close).
Tests found 12,953, creating a gain of 7% (average), with a 64% failure rate, and making $2.74 per share.
The figure on the right shows a hook reversal in an uptrend.
The pattern forms an inside day (price today has a lower high and higher low than yesterday). I added a rule that today's high can't equal today's low. You see that in four price doji (a pattern where the open=high=low=close. You see them often in thinly traded stocks).
After that, the open must be near the high and the close must be near the low.
In the figure, I show the colored price bar inside the trading range of the prior price bar. Today's open (in red) must be near the intraday high (within 25% of the candle's height, as measured down from the high price) and must be 25% of the candle's height below the intraday low. I show that in the figure with a horizontal cyan bar.
This just in! This is a rare pattern. I found 1,136 of them and they saw price drop an average of 7%, with 65% of them failing to see price drop more than 5%, and each share made an average of $2.55.
See the Below 3 pattern.
Look for price to make a high followed by the next two price bars staying below the first bar's high.
The average percentage profit tied for second place, failures at 65% were in the middle of the pack, and so was the profit per share, at $2.84.
This is one of the more complicated patterns. I show it in the figure to the right. Find two price bars in an uptrend which satisfy these rules.
In short, look for an outside day with the open above the prior close and a close below the prior low.
Tests show it's rare, with only 1,605 of them found in this test. They averaged a 7% gain, with 66% of them failing and making $3.22 a share (third).
This pattern, along with the Below 1, 2, and 3 patterns, doesn't have a formal name. I just call it a pivot because that's what Carl Vanhaesendonck calls it.
The figure on the right shows an example of a pivot.
Price makes a higher high and higher low (A) than the prior day (B) followed by price dropping below the first day (D below B). The drop at D need not occur in one day. However, for these tests, I did look for a low on price bar D to be below the low of B. Also, the high at A must be above the high at D.
When price drops below the low at B, that's the entry signal. Enter the trade at the open the next price bar.
Tests show the pattern is rare, giving traders an average of 7% with 66% failing to see price drop more than 5%, but the profit per share is high, at $3.29, placing second.
See the Below 3 pattern.
I tested this one just to see how it would compare with the Below 2 and Below 3 patterns.
Look for price to make a high and the following day to make a lower high. That's it.
Performance is last. It's plentiful (50,120 found) with each pattern averaging a drop of 6% (worst), with 69% failing (almost last), and making $2.51 per share (placing 9th).
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."
Magnet. Something acted upon by magnetism.