Japanese candlesticks — data, aggregation and pattern rules

A Japanese candle summarizes four prices from a chosen interval. It does not record participants’ thoughts or the complete path between those prices. Start with the data source, time boundaries and information available at the decision; then compare shapes under explicit rules.

Check the chart’s prices and colour rule

Open is the first recorded price in the interval, High its maximum, Low its minimum and Close the last. Body length is |Close − Open|. Upper shadow is High − max(Open,Close); lower shadow is min(Open,Close) − Low. Their sum equals High − Low for valid OHLC.

In MetaTrader 5’s documented OTC mode, candles use Bid quotes; exchange charts use Last deal prices. A Bid Low is not automatically an available buying price. Chart observations and executed order prices are separate records.

Ordinary body colouring compares Close with Open. TradingView also offers colour relative to the preceding Close, which answers a different question. For example, assigned previous Close 105 and current O 100, C 103 produce a rising body but a decline against the previous Close. Record the setting alongside symbol, feed, interval and clock before comparing charts.

Three complete M5 bars form one M15

These are invented prices in arbitrary units, not pips. All three bars use one source and clock, with aligned boundaries and complete data. The M15 covers the same time as the three M5 bars; it is not a later fourth bar.

Use the first Open, largest High, smallest Low and last Close. The aggregate is (100,106,97,98), with range 9, body 2, upper shadow 6 and lower shadow 1: 2 + 6 + 1 = 9. Do not average Opens or sum body lengths.

The individual body lengths are 3, 1 and 4, totaling 8. The aggregate body is only 2 because signed changes +3, −1 and −4 sum to −2. A larger interval preserves extrema and endpoint change while discarding part of the sequence; that does not make it a universally stronger signal.

Account for gaps between adjacent intervals

For consecutive records, last Close − first Open equals the sum of signed bodies plus the sum of each next Open − previous Close. Absolute body lengths cannot replace signed changes.

Aggregate OHLC is (100,110,100,108), change +8. Bodies cancel to zero, while the gap is 110 − 102 = 8. Thus 0 + 8 = 8. The records do not supply the observation path between the first bar’s Close and the second bar’s Open. It cannot tell you whether a real gap arose from quotes, session boundaries or missing data; that requires source records.

The separate clock diagram groups five explicitly supplied observations into [00:00,00:20) and [00:05,00:25). The intervals have equal length but different endpoints: bodies +3 and −6. Opposite colours do not by themselves prove a data error.

A current bar is not a completed observation

Just after the first two M5 bars close at 09:10, the supplied M15 data are O 100, H 106, L 99 and current C 102. Its body is positive. After the third bar completes at 09:15, final C is 98 and the body is negative. The final historical mark was not available at 09:10.

A rule requiring the completed M15 Close can use it only after that interval ends. It cannot claim an earlier entry at Low 97 based on the later completed shape. A rule for open-bar decisions needs the observations available at each decision and a separately reproducible timing model.

Pattern names organize different geometric questions

Doji describes Open and Close being close under a stated tolerance; spinning top concerns a small body relative to both shadows. Ideal full marubozu has no shadows, but definitions can include opening or closing variants. State the exact version rather than assuming every detector uses identical thresholds.

A long lower-shadow shape is described as a hammer after a decline and a hanging man after an advance. Upper-shadow counterparts also depend on context. These names do not identify actual trader intentions or certify reversal. The pin-bar discussion separates shape, context and a chosen trading rule.

Keep the complete observation set when counting a pattern. Do not search until a quota of attractive examples is met, or infer frequency from the names. Threshold, feed, clock and equality handling can all change which candles are included.

An engulfed body need not engulf the full range

For this separate illustration, require the second rising body to extend strictly beyond both endpoints of the first falling body. This is a geometric exercise; a full contextual pattern needs additional predefined conditions.

Body B starts below 100 and ends above 103, so it passes this strict body rule. Yet its Low is higher and its High lower than A’s: B’s full range is inside A’s, not outside. Harami instead compares a smaller second body contained in the first. The pattern reference addresses separate definitions; removing a required gap or changing equality creates a different tested rule.

Separate price path, shape and trade result

Both paths 100 → 106 → 97 → 98 and 100 → 97 → 106 → 98 produce the same aggregate OHLC. They do not establish the same order of threshold crossings. OHLC alone also cannot count participants or explain their motives.

Price context can be an explicit feature, such as a previously defined level or trend measure, without guaranteeing a better result. The support and resistance lesson explains reference areas. Do not invent a coin-toss success rate or claim that three confluences are independent merely because their labels differ.

Assess data validity, geometric classification and trading outcomes separately. A correctly identified shape can still precede an adverse move. Record decision time, fills, size, costs and cases without the hoped-for reaction before judging a strategy.

Reproduce the example before testing a trade

Append one hypothetical fourth M5 bar: O 98, H 101, L 96, C 100. The combined 09:00–09:20 interval is (100,106,96,100). Body is zero, range 10, upper shadow 6 and lower shadow 4. A zero endpoint change hides substantial movement inside.

Next, keep one feed and clock, specify classification rules and retain the complete chosen sample. A fixed month schedule, twenty demo trades or one hundred live trades does not certify readiness or an edge. If you need the single-candle anatomy first, read Japanese candlestick basics. Evaluate a fully specified trading hypothesis separately from this lesson’s arithmetic.