Advance Block Candlestick Pattern: Weakening Three-Candle Rallies
An advance block candlestick pattern is a three-candle warning of a weakening rally: bullish candles close successively higher while smaller bodies or longer upper shadows show less progress within each bar. It does not establish that the trend has reversed. Measure the completed three-bar shape first, then record any later downside confirmation as a separate event.
Key Takeaways
- A higher close can coexist with a smaller body: the open has moved higher too.
- Use explicit body, shadow and opening-position thresholds; detector definitions differ.
- The worked example uses a deliberately strict combination of shrinking bodies and growing upper shadows.
- A later close below candle 3 low is a chosen confirmation rule, not part of the original three candles.
Choose a definition before labeling the chart
TA-Lib describes consecutive rising closes, opens within or near the preceding body, and a long first body with a short upper shadow. Its implementation accepts several weakening combinations, including a smaller final body with a long upper shadow. It does not require both bodies and shadows to change monotonically in every accepted pattern. The stricter specification below is an educational example, not a reproduction of TA-Lib or a claim about a TradingView preset.
- Number candles 1, 2 and 3 from oldest to newest. Use completed, ordinary OHLC candles from the same symbol, timeframe and session.
- For a bullish candle, body B = close minus open, upper shadow U = high minus close, and range R = high minus low. Require B > 0 and R > 0 before calculating ratios.
- Opening position P = (current open minus previous open) / previous body. A strict inside-body open has 0 < P < 1; equality is excluded in this version.
- Keep trend context separate. One testable uptrend filter is a close above the 20-bar SMA with that SMA above its value five bars earlier, evaluated on the bar before candle 1. This is a research choice, not a universal definition.
A quantitative rule set for the weakening rally
Freeze the average absolute body of the 20 bars preceding candle 1. Require enough prior data and a positive average. These thresholds make the example reproducible; they have not been optimized or shown to improve returns. Requiring both shrinking bodies and growing shadows excludes some patterns accepted by broader definitions.
| Measurement | Chosen rule | Meaning |
|---|---|---|
| Direction | C1 > O1, C2 > O2, C3 > O3; C1 < C2 < C3 | Three bullish bodies with rising closes. |
| First candle | B1 > frozen 20-bar average; U1 / B1 <= 0.10 | An initially large body and limited upper shadow. |
| Open placement | 0 < P2 < 1 and 0 < P3 < 1 | Each later open is strictly inside the previous bullish body. |
| Body contraction | B2 / B1 <= 0.80 and B3 / B2 <= 0.80 | Each later body is at least 20% smaller. |
| Upper shadows | U1 < U2 < U3 and U3 / B3 >= 1 | Shadows grow, with the last at least as long as its body. |
Worked advance block: closes rise, bodies shrink
Assume the prior trend filter passes and the frozen average body is 2 price units. The bodies below are 4, 3 and 2: contraction ratios are 0.75 and about 0.667, or reductions of 25% and 33.3%. Upper shadows grow from 0.2 to 1 to 2.5. The last shadow/body ratio is 1.25. Opening positions are (103 - 100) / 4 = 0.75 and (105.5 - 103) / 3 = 0.833. All selected rules pass even though closes rise from 104 to 106 to 107.5.
| Candle | Open | High | Low | Close | Body / upper shadow |
|---|---|---|---|---|---|
| 1 | 100 | 104.2 | 99.8 | 104 | 4 / 0.2 |
| 2 | 103 | 107 | 102.8 | 106 | 3 / 1 |
| 3 | 105.5 | 110 | 105.3 | 107.5 | 2 / 2.5 |
| Bar | Open | High | Low | Close |
|---|---|---|---|---|
| 1: body 4 | 100 | 104.2 | 99.8 | 104 |
| 2: body 3 | 103 | 107 | 102.8 | 106 |
| 3: body 2 | 105.5 | 110 | 105.3 | 107.5 |
Hollow bodies close at or above the open. Filled bodies close below the open.
Invented data for arithmetic, not a historical trade. Prior trend and the 20-bar baseline are assumed outside the diagram. No future candle is needed to assign the shape label.
Compare a healthy three white soldiers structure
TradingView describes three white soldiers as long bullish candles opening inside the preceding body and closing near their highs. The comparison below has equal four-unit bodies and 0.2-unit upper shadows, about 4.5% of each 4.4-unit range. Later opens are 75% of the way through the prior body. It passes those geometric checks against the same assumed two-unit baseline, but fails this advance block rule because its bodies do not contract and its shadows do not grow. Healthy describes the candle structure here, not the profitability of buying it; a reversal interpretation also needs the preceding decline.
| Candle | Open | High | Low | Close | Body / upper shadow |
|---|---|---|---|---|---|
| 1 | 100 | 104.2 | 99.8 | 104 | 4 / 0.2 |
| 2 | 103 | 107.2 | 102.8 | 107 | 4 / 0.2 |
| 3 | 106 | 110.2 | 105.8 | 110 | 4 / 0.2 |
| Bar | Open | High | Low | Close |
|---|---|---|---|---|
| 1: body 4 | 100 | 104.2 | 99.8 | 104 |
| 2: body 4 | 103 | 107.2 | 102.8 | 107 |
| 3: body 4 | 106 | 110.2 | 105.8 | 110 |
Hollow bodies close at or above the open. Filled bodies close below the open.
Illustrative geometry only. Equal bodies can qualify as three white soldiers; increasing body length is not required. The diagram does not include prior trend context.
Record later downside confirmation without rewriting the label
The advance block example becomes a completed shape at candle 3 close. For a separate, illustrative follow-through test, watch the next three completed bars for the first close strictly below candle 3 low, 105.3. Cancel that pending test if a bar first closes above the fixed pattern high, 110, or if the three-bar window ends without confirmation. Keep both levels fixed. Neither the expiry period nor these price levels are a universal trading rule.
| Later observation | Result under the chosen rule | Historical label |
|---|---|---|
| Low 105.0, close 105.6 | No confirmation: only the wick crossed 105.3. | Keep the three-bar advance block label. |
| Close 105.2 within the window, before cancellation | Downside confirmation on this later bar. | Keep the original label and add a separate dated event. |
| Close 111 before downside confirmation | Cancel the pending downside test. | The valid shape remains; the bearish hypothesis failed this test. |
| Three bars finish without either closing break | Expire the pending test. | A warning without the selected follow-through. |
- A close below 107.5, the third close, is not the same as a close below 105.3, the third low. Select the level explicitly.
- Do not move a later confirmation marker back onto candle 3 or filter historical shape labels using future outcomes.
- TradingView documents that the realtime closing update confirms a bar. Waiting for that update prevents an unfinished candle from deciding this close-based rule.
- If testing a strategy, specify entry, stop, exit, costs and fill timing separately. A confirmation close is an observation, not a guaranteed execution price.
Check failures and detector disagreements
An open equal to the prior close fails this strict inside-body version. On data with consecutive opens at the preceding close, that choice can exclude many samples; test an inclusive boundary separately rather than changing it after seeing results. A tiny or zero body makes shadow/body ratios unstable or undefined. The bullish-body and data-validity checks must run before division.
- A longer upper shadow records a high that did not hold through the close. OHLC alone does not identify who traded or prove distribution.
- Volume or RSI filters add hypotheses. Compare each with the same unfiltered baseline, entry/exit rules and costs; do not assign a higher success rate without a documented sample.
- Changing session boundaries, timeframe, price adjustments or detector thresholds can change the labels. Record those choices alongside each chart.
- Measure signals, confirmed events, expired events and subsequent trade outcomes separately. Reserve later data for evaluation after selecting parameters.
Pineify next step
Turn the measured rules into an inspectable detector
Ask the Pine Script AI Coding Agent to draft the body, shadow and opening-position conditions, with separate markers for the three-bar shape and later confirmation. Review its source and compare marked candles with their OHLC values in TradingView. AI Chart Analysis can provide a second visual review of a clear, labeled screenshot; verify exact ratios from price data because an image review can miss or misread candles.
Educational information, not investment advice. All OHLC prices, thresholds and confirmation windows on this page are hypothetical research examples. A pattern does not guarantee a reversal or profit. Trading can cause substantial losses, and execution prices may differ from observed chart levels.