Hammer is a small-bodied candle near the top of its range with a long lower shadow, considered after a decline. Separate the candle’s shape, preceding context, chosen confirmation rule and actual trade. Its name describes a possible reversal candidate; it guarantees neither a rise nor an executable entry.

What the hammer looks like and why it has that shape

For a completed candle, use O for open, H for high, L for low and C for close. Body B=|C−O|, lower shadow D=min(O,C)−L, upper shadow U=H−max(O,C), range R=H−L. First validate L≤min(O,C)≤max(O,C)≤H. StockCharts describes the long shadow as generally at least twice the body. For our own shape exercise choose B>0, D≥2B, U≤0.25B and B/R≤0.30. The last two thresholds are our arbitrary filters, not universal defaults or optimal settings. With B=0, do not divide by the body; classify doji separately.

Either a rising or falling body can fit the selected shape. Colour alone does not establish the result or reveal the participants’ motives. The candlestick patterns worth knowing first guide explains body and shadows. Keep the quote side, instrument, interval and candle boundaries consistent when comparing records.

A hammer shape cannot repair inconsistent OHLC

A long lower range may attract attention, but the first check should be record consistency. Pattern measurements require sound data.

Scroll sideways to see the full diagram

A hammer shape cannot repair inconsistent OHLCA flawed fictional EUR/USD record has Open1.1000, High1.1008, Low1.0970 and Close1.1010. The navy segment shows assigned Low–High on a linear price axis. The coral Close lies0.0002, or2 pips, above High. This illustrates contradictory data, rather than a valid candle or Hammer.Check data before naming the candleFlawed fictional EUR/USD record.1.10001.0970Close = 1.1010High = 1.1008Open = 1.1000Low = 1.0970Close>High by2 pips: inconsistent.Pause pattern assessment; check source.

Our data counterexample: EUR/USD, pip=0.0001; Open1.1000, High1.1008, Low1.0970, Close1.1010. The record is deliberately erroneous. Close−High=0.0002, or2 pips. The diagram shows Low–High and a separate inconsistent Close point, rather than a valid candle.

OANDA’s OHLC definition identifies High as the highest, Low the lowest and Close the last price in the same period. Open and Close in a consistent series using the same quote side must therefore lie within Low–High.

Our checking procedure: flag the record, check its period and quote side, and retrieve the correct source again. Do not replace High with Close merely to make the drawing consistent. Without source data you cannot identify which field is wrong. This checks data quality; it identifies no Hammer and makes no forecast.

Close above High is a contradiction to resolve before classifying the candle.

What the four prices show, and what they leave unknown

Our hypothetical EUR/USD candle has O=1.1000, H=1.1012, L=1.0970 and C=1.1010; one pip=0.0001. Its body is 10 pips, lower shadow 30, upper shadow 2 and range 42. D/B=3, U/B=0.20 and B/R≈0.2381 pass our three ratio filters. Change only C to 1.1002: body 2 pips, lower shadow 30 and upper shadow 10. D/B=15 and B/R≈0.0476 still pass, but U/B=5 fails. A very long lower shadow alone does not satisfy every condition.

OHLC shows that the first candle closed above its low. It cannot establish the identities, motives or entire sequence of transactions. O→H→L→C and O→L→H→C can yield identical OHLC. If both a stop and target lie within one candle, its high and low alone cannot determine which occurred first. TradingView documents its emulator’s assumptions and lower-timeframe refinement; do not silently choose a favourable path.

One candle is insufficient to reproduce every detector. TA-Lib C 0.6.4’s default CDLHAMMER also compares its body with the previous ten. Keep a hypothetical candidate O/H/L/C=90/100.5/60/100 unchanged: B=10, D=30, U=0.5. Give nine preceding candles O/H/L=100/115/90 and C=105 or 112; the immediate predecessor is 100/102/90/95 in both trials. Add three older candles of the same type for warm-up. Only those earlier closes change.

Our fixed candidate with two different histories
Previous ten bodiesMean bodyOur shape filtersCDLHAMMER code
Nine bodies of 5; last body 55Pass0
Nine bodies of 12; last body 511.3Pass100

Body 10 is not below 5 but is below 11.3. The other default conditions pass in both trials: lower 30>10, upper 0.5<2.37 and body bottom 90≤previous low 90+Near 5. Code 100 is a classification, not a 100% chance or evidence of a better trade. Preserve earlier OHLC as well as a screenshot. The existing saved 120-bar Kraken EUR/USD H1 sample, 21–25 September 2026, contains five nonzero CDLHAMMER codes; this count classifies one sample and measures no profitability.

The inverted hammer and the hanging man — same shape, different place

The inverted hammer and shooting star comparison changes which shadow is long. Inverted hammer is considered after a decline with the long upper shadow, while shooting star is considered after an advance. These descriptions do not establish that one pattern is inherently weaker, demands stronger confirmation or predicts a specific next move.

Hammer and hanging man can have identical small bodies and long lower shadows. The former is considered after a decline, the latter after an advance. A context rule is therefore needed alongside shape. Naming two contexts does not measure their reliability or justify importing an unverified stock-market continuation statistic into forex.

How to trade the hammer step by step

Define preceding movement before evaluating outcomes. Our simple context exercise requires three successively lower closes, such as 1.1080, 1.1050 and 1.1020 before the candidate. This is our rule, not a complete theory of trend. TradingView’s Hammer tool allows SMA50, SMA50/SMA200 or no trend detection, so scanners can disagree. An optional confirmation rule requires the next completed close above the hammer’s H: 1.1015 passes H 1.1012. Only then is confirmation known. Do not backdate an entry to 1.1010 or assume an available Ask equal to a Bid candle’s close.

Confirmation changes the decision time and possible price; it does not prove buyers control the market. A stop below the candle’s low can be one chosen exit rule, but the pattern’s name does not impose it or a nearest-resistance target. The support and resistance guide provides context for defining levels. Recalculate position size and costs for the actual selected entry and exits; a longer shadow can enlarge a chosen stop distance without strengthening a strategy.

A separate fictional transaction assumes an actual purchase of 10,000 EUR at Ask 1.1018 and sale at the Bid prices below. Ordinary stop level 1.0968 and target 1.1118 are arbitrary model inputs. Total commission is 2 USD; omit financing and conversion. These assumptions do not recommend levels.

Assumed full exits of our 10,000 EUR position
Sale BidPrice P&L, USDAfter commission, USD
1.1118: target+100+98
1.0968: at stop level−50−52
1.0958: worse execution−60−62

Execution prices already contain spread. Reward:risk after commission is 98/52≈1.88, whereas the price-only ratio 100/50=2. Ordinary stops can execute worse; IG’s UK documentation distinguishes guaranteed stops as a separate product feature. This table states no maximum possible loss. Keep signals and actual fills separate, including missed entries, absent confirmations and continued declines.

A daily candle is not automatically more reliable than an hourly one. A different interval changes aggregation and the test. The RSI indicator can supply a separately defined filter, but indicators calculated from the same prices are not independent evidence merely because their names differ. Evaluate fixed rules on separate later data rather than treating confluence as proof.

The most common mistakes when trading the hammer

Keep shape, context and confirmation definitions fixed before looking at later prices. A shape outside your chosen decline filter can fail that model while another scanner still labels it. Record every candidate, including no confirmation, continued declines and later returns below the low. A successful-looking historical selection does not measure a strategy’s false-alarm rate.

Stop placement, entry timing and colour need explicit rules rather than blanket instructions. A stop halfway up a shadow is not guaranteed to be hit, and a stop below it is not guaranteed to cap loss. Compare the separately defined engulfing pattern and morning star without assuming all detectors or confirmation rules coincide. Costs, quote sides and information available at the decision time remain essential.

What to do tomorrow

  1. Write a reproducible shape and context definition. Keep candle boundaries, quote side and preceding OHLC, including records rejected by the model.
  2. Record when any confirmation becomes known. Choose entry, exit and position-size rules separately and retain actual or simulated execution assumptions.
  3. Compare filters on unchanged later data, including costs and missed fills. Two price-derived clues are not automatically independent evidence.
  4. Use demo to practise the procedure. One week or a few successful candles alone demonstrates neither a statistical edge nor readiness to risk live money.

If you want to organise all your knowledge of candlesticks within a broader course, a good starting point is the technical analysis section in our library.