Pattern Reference

See all 19 pattern types with labelled charts and price examples.

Every Top Ranked Names entry on Signal Desk has a chart below it, for Daily and Weekly boards. Charts load as you scroll and can be enlarged. Each chart uses the saved signal and stops at its selected date, including on older boards. Price charts keep the price axis at or above zero. When the full price range is at least eight times the recent 30-candle range, chart cards show one chart at the usual size with a Recent 30 / Full chart switch. The recent view shows price candles only. The full chart keeps its pattern lines and moving averages. Enlarged charts and Research Pages show both views together, side by side or stacked on a phone. Both views stop at the same selected date. This display change does not alter scores, trade rules, or the images used for AI reviews. If chart data is unavailable, the entry keeps its Research Page link and offers a retry. Fibonacci charts use regular quarter labels and preserve real calendar gaps. Current and past scan downloads include the saved Fibonacci level, its price, the main move dates and prices, and the reaction date. History coverage is checked for each ticker. Partial ticker histories do not inherit a complete status from the market.

Backtests show past results. They add no points to the current Composite score.

Saved and Current Signals

Each saved Top Trades category uses the exact stored report for its date and timeframe. Source data and reviews stay linked to that report. Archive files are checked against those reports before publication.

Combined score and trade plan

1. All Top Trades use one fixed 100-point score. A current Daily signal is not required. Weekly-only ideas can qualify.
2. Use current Daily and latest completed Weekly signals. Daily prices stop at the saved market date. Missing Daily patterns earn no Daily points. No unfinished weeks or later data.
3. Take five per category. Top AI reviews the combined top 10. Show the five highest AI-adjusted scores using chart reviews only. Fundamental and macro reviews are disabled. Liquidity compares ideas within each market. Combined Top Trades has no Daily/Weekly switch. Charts and Signal Desk keep it.
4. Each combined category shows its own saved Top 5. Its order stays the same in the board, Full Report, and CSV.
5. Historical boards with only saved Weekly scans use those signals and Daily prices through the board date. Missing Daily scans add no Daily signal points. Daily price dates do not count as completed Daily scans.
6. Saved AI reviews must match the stock, direction, signal evidence and prices for that board date. Show the chart evidence date. Missing reviews leave blank scores and visible gaps. Top AI orders the same five Composite candidates by their chart-only AI scores. Opening a historical board does not add Journal trades.
7. For historical Top Composite and Top AI boards filtered against the Journal: use the journal's trade dates. Exclude a stock while it is open in either direction. Keep its first planned day only. After it closes, require a new main setup formed after the exit. A new scan date, score or supporting pattern alone does not count. Skipped trades do not block a stock. Fill each space with the next eligible saved candidate. Keep saved scores. Top AI ranks those same five candidates using their dated chart reviews. This board update adds no journal trades.
8. Composite scores use a 0–100 scale. A higher score means stronger support under the current method. It is not a win probability or an expected return.
9. The score considers pattern quality, trend direction, and agreement between candle patterns and chart structures. It uses Daily and completed Weekly evidence.
10. Fibonacci uses only 38.2%, 50%, and 61.8%. The 38.2% and 61.8% levels carry equal weight, with less weight at 50%. Only a confirmed price reaction after a significant main move can add to Composite. Fibonacci shares the existing chart-structure limit. The same price boundary counts once.
11. Conflicting signals reduce the score. Repeated descriptions of the same setup do not add separate confirmation.
12. Backtest results do not add points to the current Composite score. Scoring needs 50 consecutive completed Daily sessions.
13. Journal entry needs a final score above 60. The above-60 rule applies only to journal entry.
14. New AI trades can enter during the next regular trading session only. Use the open if suitable; otherwise use the 0.5 ATR limit from the signal close when touched. An unfilled order expires at that session's close. Confirm simulated results from completed daily prices. Freeze ATR14: mean of 14 true ranges through the signal date.
15. On a later entry's first day, ignore the earlier opening price for exits. If stop and target both touch, count the stop first. A target-only touch needs a close at or beyond the target; otherwise the price order and return remain unresolved.
16. Trades recalculated after a rule change are historical simulations. Their trade details and CSV identify the recalculation date. Original records remain saved.
17. Provider rounding in refreshed adjusted prices does not change saved entries. Larger price differences or changed entry conditions need review.
18. Full stop and target: each 2 ATR from entry. There is no time limit. No partials or trailing stop. Enabled trades also exit on confirmed setup failure. A repeated signal does not change the plan.
19. Opening gaps fill at the open. If both levels touch, stop first. Missing prices remain unresolved. One open position per stock.
20. Saved prices refresh the AI journal before new trade selections and before each market publication. Entries and exits follow their saved rules.
21. AI Journal totals include only trades admitted under the current Journal rules, including Weekly-led ideas. Recalculating a board does not add past trades. All Journal returns exclude trading costs. Totals add trade percentages; personal trades keep saved holding periods.
22. For new journal entries, the review must exist before entry. Late reviews are skipped.

Historical scores

The pattern score measures one detected pattern. The Composite score combines the setup evidence. The AI score adds a chart review adjustment. The liquidity score also considers trading activity. These scores measure different things. None is a chance of profit.

Views

AI Verdicts

• Pro can download the completed AI Ticker Review and its charts as a PDF.
• Downloads use the saved review, not another AI request, and remain available for seven days.
• Completed Daily/Weekly charts produce a verdict, direction, levels, and risks.
• Use Composite's direction rule before grouping signals, calculating the candidate score, and sending evidence to the AI.
• A held break above a double or triple top is BUY.
• A held break below a double or triple bottom is SELL.
• Broken trendlines and channels follow the confirmed break. Range-box signals use their support, resistance, or held-break state.
• Detected Scanner Patterns shows these directions. A BUY pattern does not by itself make the full review a BUY.
• Saved reviews keep their original results.
• Existing setups receive an AI score adjustment; other charts receive a separate 0–100 score.
• AI Ticker Review and the free example use Top Trades' current combined Composite formula before AI adjustment.
• Same dated prices, eligible signals, and direction give the same starting score. Backtests add no points.
• Composite scoring needs 50 consecutive completed Daily sessions. Missing inputs stop the request before AI runs.
• Chart analysis only. Uses your plan quota; no automatic journal entry or scanner-rank change.
• New technical reviews hide tickers and company names. Chart evidence stops at the saved cutoff date.
• Later bars and pattern dates are rejected. Previous AI answers and trade outcome fields are removed from technical inputs.
• Top AI boards show chart reviews only. Fundamental and macro reviews are disabled.
• Each new review saves its dated input, prompt, schema, and rule version. Later outcomes do not change its score.
• Review inputs enforce the saved cutoff. Later bars, outcomes, and revised trade results do not enter the score.
• Rule testing uses a fixed later date window. Its results cannot be used to change that window's rules.

The journal keeps shared AI entry and exit rules in AI Trade Methodology. It shows why a trade was skipped next to its status. Open the trade details for AI chart notes and any missing-price details. My Tracked Trades has separate holding-period rules. Both use returns before trading costs.

• Scheduled Journal candidates come from the 10 AI-reviewed setups. Top AI displays five. Any of the 10 can qualify with a final score above 60 and all other entry checks passed. A Weak or Reject verdict alone does not prevent entry.
• Plans use saved dated evidence and fixed trade rules. Returns use prices after entry. The page and CSV show signal, entry, and exit dates.
• Every Trade Journal list and CSV download shows the newest signal date first, including Open Trades, Closed Trades, AI Trades, My Tracked Trades, and Latest 5 Trades. Entry dates, exit dates, time added, state, and timeframe do not change this order. Missing signal dates stay blank and sort last.
• AI trades use Daily entry rules. Daily and Weekly patterns can both add to the score. Totals include only trades admitted under the current Journal rules. Other records remain private. Recalculating a board does not add past trades or change recorded returns.
• Require a final score strictly above 60 under the current method.
• New AI trades can enter during the next regular trading session only. Use the open if suitable; otherwise use the 0.5 ATR limit from the signal close when touched. An unfilled order expires at that session's close. Confirm simulated results from completed daily prices.
• Save both the AI review and selection before that open. Late reviews or selections are skipped. One active position per stock.
• On a later entry's first day, ignore the earlier opening price for exits. If stop and target both touch, count the stop first. A target-only touch needs a close at or beyond the target; otherwise the price order and return remain unresolved.
• Trades recalculated after a rule change are historical simulations. Their trade details and CSV identify the recalculation date. Original records remain saved.
• Provider rounding in refreshed adjusted prices does not change saved entries. Larger price differences or changed entry conditions need review.
• ATR14 is the average daily price range, including gaps, over 14 completed trading sessions through the signal date.
• Full stop and full target are each 2 ATR from entry. Freeze ATR14 at the signal date. There is no time limit. Enabled trades can also exit on confirmed setup failure. See Daily setup checks below.
• No partial exits or trailing stop. Keep the levels fixed. Gaps fill at the open. If both levels touch, stop first.
• Closed trades use their exits; open trades use the latest saved price in place of the exit price.
• Average closed-trade return, total, high, low and win rate use closed results; wins exceed 0%.
• The Open Trades view keeps current returns separate.
• Signal Desk shows latest saved US/HK AI results, regardless of board date, with the AI Journal's access delay. Personal trades are excluded.
• Results are simulations before trading costs.
• Buy return (%) = (exit price − entry price) ÷ entry price × 100.
• Sell return (%) = (entry price − exit price) ÷ entry price × 100.
• All Journal returns exclude trading costs, including fees and slippage (the gap between expected and actual trade prices).
• Cards and charts keep the sum of trade returns in percentage points (pp). Average return is a percentage. Charts are on the main page; notes: Methodology.
• The average uses only trades with an available return. Its count shows those trades. Average return = sum of those returns ÷ that count.
• Gains and losses both count. A 0% return counts. Planned and skipped trades do not enter the total or average.
• Trades waiting to start and trades with no return available appear separately. Missing prices remain unresolved; a missing return is not 0%.
• The sum gives each trade equal weight. It does not compound returns or account for position sizes. It is not an overall account return.
• Example: +6%, 0%, and −3% give a total of +3 percentage points and an average of +1% from three trades.
• The AI summary has one total. The sum is not a portfolio return or a fixed-dollar-risk result. Source, market, and status views keep the same format. Only My Tracked Trades shows a Daily versus Weekly period comparison.
• Pro can download a CSV of the selected view. Its returns match the page. Personal tracked trades stay private to each account.
• Personal trades enter at the first stored regular open after saving. An older signal never creates an earlier entry. They can be deleted individually or cleared together.
• Personal tracked trades use their saved holding periods and close after that many stored trading sessions. The entry session counts as one.

The journal shows one review card for each open AI trade with daily checks enabled. Each card explains the effect on the journal position and shows the last saved check date. Warnings and scheduled exit dates stay visible. Open Why this status for the price trigger, trend warning and past checks. Sloping Daily lines follow the original slope. Horizontal and candle levels stay fixed.

No exit signal means the latest review did not trigger an early exit. Warning means more daily closes must confirm the warning. Exit scheduled shows the market opening chosen for a full journal exit. Check incomplete means required prices are missing; old counts are not a new all-clear. Review pending means the next daily review is not complete yet. These labels are not new AI scores.

A daily close is the price at the end of a trading day. Two closes in a row below the day's exit trigger for Long, or above it for Short, confirm a price exit. A move during the day does not count for this rule. A close back within the limit resets the count; a close exactly at the trigger does not count.

The trend reversal warning starts only after the trend first supports the trade. Not active yet is not an exit signal. Once active, three consecutive days must meet the opposite-trend rule. The 20-day and 50-day averages are average closing prices over those trading days. A scheduled exit stays scheduled if prices later go missing. Stop and target rules still apply. These checks do not place orders with your broker.

• Check every enabled open AI trade after each market close, even after it leaves the Top 5. This is a rule-based check, not a new AI score.
• Save one Daily setup level. Use the counted primary Daily price boundary; otherwise use the highest-point counted Daily boundary. If neither is available, use the signal-day low for Long or high for Short. Show which source was used.
• Extend the original Daily trendline or channel edge by its saved slope for each market session. Keep its original points and slope; never fit a new line. Horizontal levels, Fibonacci retracement levels and signal-candle highs or lows stay fixed. If the original slope is missing, keep the saved fixed level and label that limit.
• Keep the signal-date ATR and the 0.25 ATR buffer fixed. Two consecutive Daily closes more than 0.25 ATR beyond that day's level against the trade confirm a setup failure. Compare each close with its own day's failure price. A close back within the limit resets the count. One close beyond the setup level is a warning.
• The trend check starts only after close > MA20 > MA50 for Long, or close < MA20 < MA50 for Short. Then three consecutive closes with the full opposite ordering confirm a trend failure. MA20 and MA50 are the mean of the last 20 and 50 completed Daily closes.
• A confirmed failure queues a full exit at the first regular-session opening after the decision is saved. A late check cannot claim an earlier opening. Show the decision date and the actual exit date separately.
• The fixed 2 ATR stop and target remain active. An earlier stop or target takes priority. At the queued opening, a gap through either level keeps its stop or target label. There is no fixed holding deadline, no partial exit, and no wider stop.
• Only completed prices available at the check can be used. Missing required prices leave the review incomplete. No missing price is treated as a failure or a pass. An exit already queued stays queued while prices are missing.
• Refresh Journal checks before release. Block old checks if prices exist. Show dated gaps as incomplete reviews.
• The live site must confirm the same file and current Journal checks. Retry pending builds without repeating completed calculations.
• Each trade shows its setup level, status, checked-through date, confirmation counts, and saved check history. Original entry scores stay fixed. A lower score, loss of Top 5 rank, or an expired scanner label does not cause an exit.
• Setup checks begin only after they are enabled for a trade. Completed trades and earlier exits keep their recorded results. Current checks never create a past exit.
• A line-rule change starts with the next uncompleted Daily session and resets only the setup-level close count. Earlier saved checks keep the price and rule used then. The trend count, original entry, hard stop and target stay unchanged. An already queued exit is not cancelled.
• Daily rules control these automatic journal exits. AI Ticker Review remains an optional separate chart opinion. It does not cancel a confirmed exit. These are journal calculations; the website does not place broker orders.

Fixed-level example: a Long setup level is 100 and its fixed ATR is 4. Two closes below 99 confirm failure. A close at 99 does not count. For a rising line at 100 today and 101 next session, the failure prices are 99 and 100. Each close uses its own day's price. The full trade exits at a later opening, even if the stop is farther away. ATR means average true price range, including gaps.

These exit rules have not been shown to improve returns. The journal records both winning and losing exits. Open AI Trade Journal.

• Backtests measure individual patterns over fixed holding periods. AI Journal trades use AI selection and separate exit rules.
• For example, a 10-session Daily backtest measures the return after 10 trading days. An AI Journal trade can close earlier or later.
• Pattern types count once across US/HK and Daily/Weekly. Strict/Loose, setup states, and MA100/MA200 results are separate studies, not extra types.
• Stop candidates with fewer than three valid prices. Resume rejects changed inputs; saved values keep full precision.
• Reuse historical trendline checks only if prices, security details and calculation rules match. Trade returns are checked again.
• From January 1, 2010, or the ticker's later first available trading date; patterns may need additional history.
• Entry is the next market-session open after the signal. Exits: 1, 3, 5, or 10 completed Daily/Weekly periods.
• Exit: closing price at the chosen holding period.
• Buy return (%) = (exit price − entry price) ÷ entry price × 100.
• Sell return (%) = (entry price − exit price) ÷ entry price × 100.
• Positive means profit. Fees and slippage are excluded.
• Each setup, direction, quality, period, and MA100/MA200 study stays separate. Averages use full matching history.
• N counts all trades in the study. Each holding period has its own completed-result count, n.
• Average return is the sum of completed returns divided by n. Win rate is the share above 0%.
• Zero returns count in n but are not wins. Missing or pending returns do not count in n.
• Open a pattern for dated returns. Pending results are excluded; zero returns count as Flat.
• History includes archived signals and uses newer completed returns when available.
• Pro includes summary CSV downloads by market, timeframe and family. Backtest History has each ticker's trade CSV.
• CSV holding periods use sessions for Daily and weeks for Weekly. Missing outcomes stay blank; trade status shows pending or not stored.
• Charts show less history than backtests. Missing prices are not zero.
• The current ticker universe can bias results by omitting failed or delisted securities.