High-wave candle: what it means, TA-Lib rules and test results

A high-wave candle has a short real body and very long shadows reaching far above and far below it. The session swung hard in both directions and ended near where it started. In candlestick tradition the name evokes a choppy sea: big waves, no progress.

Nison describes these candles as a sign that the market has lost its sense of direction. That is a more specific claim than "indecision". A quiet small-bodied day is indecisive; a high-wave day is violently indecisive, with both buyers and sellers able to move the price a long way and neither able to hold it.

This page covers the anatomy, how TA-Lib's CDLHIGHWAVE turns "very long" into a number, why the detector's output is signed by body color and what that does in Finabase's scanner and gauge, two very different real cases from IBM and Intel in 2026, and a test of 3,876 detections.

Anatomy of a high-wave candle

What makes the shape:

1. A short real body, white or black, smaller than recent bodies. It may be a doji, but does not have to be.
2. A long upper shadow, far longer than the body.
3. A long lower shadow, also far longer than the body.
4. Usually a wide range. In the classical picture, the shadows are long not just compared with the body but compared with the surrounding candles.

When the body is reduced to a line and sits in the middle, the high wave overlaps with the Long-legged doji: rules, the TA-Lib quirk and real results and the Rickshaw man candlestick: the centered doji, rules and results. In our data 2,670 candles were both a high wave and a rickshaw man.

The exact rules Finabase uses

Finabase marks high-wave candles with TA-Lib's CDLHIGHWAVE on daily bars:

The interesting setting is "very long". TA-Lib's ShadowVeryLong has an averaging period of zero and a factor of 2, which means each shadow is compared with this candle's own body and must be more than twice as long. It is not compared with recent ranges at all.

That has a predictable effect. The smaller the body, the easier the shadow test becomes. A candle with a 10-cent body needs only 21-cent wicks to be a "high wave", even if its whole range is narrower than the days around it. The classical idea of a wide, storm-tossed session is therefore not enforced: TA-Lib's high wave is really "a spinning top whose shadows are at least twice its body".

In our sample 3,876 of the 6,572 spinning tops, 59%, also passed the high-wave test. Only 135 of the measured high waves had a range at least 1.5 times the average of the previous ten sessions.

What the long shadows say

Read the shadows as two failed attempts. At one point in the session buyers carried the price far above the open and could not keep it there. At another, sellers carried it far below and could not keep it there either. The close, near the open, is where the two efforts cancelled out.

That is different from a quiet small-bodied day. A wide high wave says large orders were hitting the market from both sides. Traders often see one at the end of an extended, fast move, when early participants are taking profits, latecomers are still chasing, and nobody agrees on value. In that setting the candle is a sign that the trend's orderly phase may be over, which can mean a reversal, a sideways range or, sometimes, a violent continuation.

A narrow high wave, which TA-Lib also flags, tells that story much more weakly. It is a small candle with proportionally long wicks, and the story is in the proportions, not the size.

Confirmation

Because the candle is directionless, traders wait for price to leave its range:

• A close beyond the high or the low. After a wide high wave the range is large, so this can take time and the eventual break can be a big move.
• Several high waves or spinning tops in a cluster, which usually describes a market building a range rather than reversing.
• A range filter: comparing the candle's high–low range with recent sessions to separate the genuinely wide ones from the merely proportional ones.

Two real high-wave candles

IBM, June 1, 2026. This is the textbook storm. IBM had gained 46.1% in ten sessions. On June 1 it opened at $322.55, fell as low as $308.00, rose to $327.98 and closed at $320.42. The black body was $2.13; the shadows were $5.43 above and $12.42 below. The range, about $20, was 1.8 times the recent average, and volume was 3.6 times its 20-day average. TA-Lib returned −100, so Finabase listed it as Bearish. The next session went the other way and closed at $329.23, above the high. A reader treating the next close as the deciding vote would have abandoned the bearish idea. But the advance did not last: five sessions after the high wave IBM was 12.4% lower, ten sessions after it 16.1% lower, having traded as low as $264.80. The candle correctly marked a market that had lost its footing; the confirmation rule pointed the wrong way first.

Intel, April 17, 2026. INTC had gained 36.0% in ten sessions. On April 17 it opened at $68.85 and closed at $68.50 with a high of $70.33 and a low of $67.74, a 35-cent black body with a $1.48 upper shadow and a 76-cent lower shadow, just over twice the body. Its range was smaller than the recent average: a high wave by proportion, not by size. Finabase would have listed it as Bearish, and the next session closed at $65.70, below the low, which a bearish reader would call confirmation. Five sessions later INTC was 20.5% above the high wave's close; ten sessions later 45.4% above, after trading as high as $100.45. A confirmed bearish label on a narrow high wave meant nothing in a stock with that much momentum.

What the data says

We ran CDLHIGHWAVE on daily bars from October 30, 2024 to October 29, 2026 for 60 large US stocks and ETFs. It fired on all 60 symbols: 2,014 white (+100) and 1,862 black (−100). Each row counts moves in the direction of Finabase's label: higher for white, lower for black. Rows with an asterisk are our own cuts of the same data.

The return column is the raw price change, so a positive figure after a black high wave means prices on average rose. Separately, the 135 measured high waves with a range of at least 1.5 times the recent average were higher five sessions later 55.6% of the time, with both colors combined.

The plain reading:

• Body color carried no forecast. Both colors finished within a point of the baseline for their direction.
• The classical setting, a black high wave after an advance, did a little better than average, 48.5% lower against 46.7%. With 987 cases that difference is a bit more than one standard error, so it is not convincing on its own.
• Waiting for a confirming close did not help either color.
• Wide high waves did not lean bearish. If anything, the 135 widest were followed by gains slightly more often than average, but that group is small.

High-wave candle versus its neighbours

The high wave sits between the Spinning top candlestick: meaning, color quirk and real results, which it is a stricter version of, and the doji family, which it overlaps without requiring a doji body. 3,181 of the 4,119 TA-Lib Doji candlestick: meaning, exact rules and what the data shows candles were also high waves.

An educational outline

This is a way to study the pattern, not a recommendation to trade.

1. Check the size, not just the shape. Keep high waves whose range is clearly wider than recent sessions; the narrow ones are proportion, not event.
2. Look at what came before. A high wave after a fast, extended run is the classical setup. In a range it means little.
3. Expect a wide stop. The candle's own range is the natural risk boundary, and on a genuine high wave it is large. Size positions accordingly or stand aside.
4. Give the break time. As the IBM case shows, the first close outside the range can be a false start.

How Finabase shows high-wave candles

• Chart markers. The High Wave Candle study marks every CDLHIGHWAVE detection. On a typical large stock that is roughly one bar in eight, so compare each mark's range with its neighbours before paying attention to it.
• Scanner: split by body color. The candlestick scanner sorts detections by the sign of TA-Lib's output, so white high waves appear in the Bullish list (S&P 500, Bullish) and black ones in the Bearish list (S&P 500, Bearish). TA-Lib's source says the sign does not mean bullish or bearish; the scanner uses it anyway because it has no other way to place a two-sided pattern.
• Signals-tab gauge. On the Signals tab, each of the 61 patterns votes from its most recent detection within the last eight bars. The high wave's vote is the color of its latest occurrence: in our data Buy on 34.5% of stock-days and Sell on 32.5%. Because every high wave is also a spinning top, the two patterns usually vote the same way on the same candle.

Frequently asked questions

What does a high-wave candle mean?

It means the session moved far above and far below its open and closed near it. After a strong trend it is read as a sign that the market has lost its direction. It does not say which way the next move will go.

Is a high-wave candle bearish?

Not inherently. Finabase lists black ones as bearish and white ones as bullish because TA-Lib signs its output by body color. In our test neither color predicted the following week.

How is a high wave different from a spinning top?

Both have a short body and shadows on both sides. In TA-Lib a spinning top needs each shadow longer than the body; a high wave needs each shadow more than twice the body. Every high wave is a spinning top, but only 59% of spinning tops are high waves.

How is a high wave different from a long-legged doji?

A long-legged doji needs a doji body, essentially no body at all. A high wave allows a small visible body, as long as both shadows are more than twice its length.

Summary

The high-wave candle is a short body flanked by long shadows, a session of large swings and no result. TA-Lib defines "long" as more than twice the candle's own body, which makes its high wave a stricter spinning top rather than a guaranteed wide-range event, and it labels each one +100 or −100 by body color. Finabase lists and counts them accordingly. Across 3,876 detections, neither color, nor the classical after-a-rally setting, beat a random day by a convincing margin. The wide ones, like IBM's in June 2026, are worth noticing as a sign of a disordered market; the narrow ones mostly are not.