Trợ giúp

Patterns scoring profile

Ranks detected chart patterns by quality, freshness, distance to price, and supporting context.

Disclaimer and responsibility

  • Scoring profiles are educational ranking tools, not financial advice and not a promise of profit.
  • Score helps decide which coins deserve manual review first, but it is not an automatic buy/sell signal.
  • Every trading decision remains the trader responsibility: position size, leverage, stop placement, exchange risk, liquidation risk, and the consequences of each trade.
  • A high Score can appear before a failed breakout, a false impulse, a weak pattern, or a news spike. Always check the chart, liquidity, higher-timeframe context, and risk before acting.

What the profile is for

  • Use this profile when automatic chart-pattern detection is enabled and the task is to decide which detected patterns deserve manual review first.
  • The profile does not treat a pattern as sufficient by itself. It also checks volume, volatility, current movement, levels, large orders, and trade activity.
  • It is useful when many patterns are found and visual review of every symbol would be too slow.

Components and weights

  • Pattern, weight 40%: detected pattern quality, freshness, and distance to current price.
  • VR, weight 15%: relative volume around the pattern context.
  • NATR, weight 10%: volatility suitable for the selected timeframe.
  • PΔ, weight 10%: current candle movement after or near the pattern.
  • Level 1h, weight 2%: whether the pattern is supported by a nearby 1h level.
  • Level 4h, weight 3%: whether the pattern is supported by a nearby 4h level.
  • Level 1d, weight 5%: whether the pattern is supported by a nearby 1d level. Higher-timeframe context receives more weight.
  • Large orders, weight 10%: whether orderbook context supports or challenges the idea.
  • TR, weight 5%: trade activity around the pattern.

Recommended settings

  • Patterns: all types enabled, minimum score 70.
  • VR: 1h, 100 candles.
  • NATR: 1h, 14 candles.
  • PΔ: 1h.
  • Levels: 1h, 4h, and 1d as separate context components.
  • Large orders: Auto threshold, from 2.5x average order size, 3-minute hold time.
  • TR: 1h, 100 windows.

How to read the score

  • High Score means a detected pattern is relatively clean and has supporting context.
  • The pattern component has the largest weight, but supporting components are important because many chart patterns fail without volume or level context.
  • If the pattern component is missing, the profile has no main signal and should not be used for that coin.
  • If the pattern is old or price has moved far away, the score should be treated as stale even when the pattern itself looks good.

What to check manually

  • Visually inspect the pattern before trusting the ranking.
  • Check whether the pattern is fresh and whether price is still near the actionable zone.
  • Check whether a level, trend, volume expansion, or orderbook density supports the pattern.
  • Check where the pattern idea is invalidated. A pattern without a clear invalidation point is not a complete setup.

Weak conditions

  • The pattern is small, visually unclear, or built from noisy candles.
  • The pattern exists without volume, level, trend, or orderbook context.
  • Price has already completed most of the expected move.
  • A lower-timeframe pattern conflicts with strong higher-timeframe context.