24h Change
24h Change helps you quickly sort coins by recent momentum. Green values mean price is up versus 24 hours ago; red values mean price is down.
This is Binance USD-M perpetual futures' rolling 24-hour price change percentage (same convention as the futures exchange ticker), not spot and not a calendar-day change. It updates continuously from the live futures ticker stream.
- Source: Binance USD-M Perpetual Futures ticker (priceChangePercent) — not spot.
- Update frequency: real-time via WebSocket as Binance publishes futures ticker updates.
To Support
Negative values mean price is below support (support is above current price). Positive values mean price is above support.
Support comes from the detected pattern's lower trendline or structural floor, stored when the pattern was last analyzed. The percentage is recalculated on every live price tick.
- support_level: from pattern detection output for that coin, timeframe, and pattern type.
- current_price: latest Binance futures mark/last price from our live kline stream.
To Resistance
Negative values mean price is below resistance (room to run up toward it). Positive values mean price has moved above resistance.
Resistance comes from the pattern's upper trendline or structural ceiling. Coins with small absolute values are near that level.
- resistance_level: from pattern detection output for that coin, timeframe, and pattern type.
- Updated live as price moves; levels themselves refresh on the next pattern analysis run.
Forming
Forming is Yes while closes stay inside the pattern's support and resistance envelopes (within 0.5×NATR). Trendlines still extend to the last closed candle.
The first close 0.5×NATR above resistance or below support starts recent breakout or recent breakdown. Forming becomes No. Trendlines freeze at the last in-band candle and are not extended. After 20 closed candles the pattern is dropped from lists.
Non-trendline patterns (head and shoulders, doubles, cups, RSI divergences) show None. Forming only applies to trendline patterns. Head and shoulders, doubles, triples, and cups stay listed while the pattern-end/break index is within the last 20 candles — the same wait as a trendline after invalidation. RSI divergences keep a shorter pivot-recency rule.
- Yes = still forming (in-band); trendlines may extend on each closed candle.
- No = recent breakout or recent breakdown; first 20 candles after invalidation.
- None = not a trendline pattern; this column does not apply.
- After 20 candles the trendline row is removed. No further data is kept.
- H&S / doubles / cups: listed while pattern end is within the last 20 candles. RSI divergences unchanged.
Fuel
Fuel summarizes volume, momentum, and leverage pressure into a single badge and zone (Low / Neutral / High) so you can compare coins at a glance in tables.
The badge maps the composite score to a label: Dead Water (low), Organic Flow (neutral), or The Powder Keg (high). Open a coin's detail page for the full gauge breakdown.
- Composite weights: Volume 30%, Momentum 25%, Leverage 45%.
- Fallback without leverage data: Volume 50%, Momentum 50%.
- Full formulas: see the Fuel Index section in this Help Center.
RSI (14)
RSI measures recent gains vs losses on a 0–100 scale using a 14-period lookback on closed candles. Values above 70 are often considered overbought; below 30 oversold — context matters by market and timeframe.
RSI is also used for divergence detection and as an input to the Fuel Momentum score.
- Period: 14 (standard RSI setting).
- Source: calculated from Binance kline closes, stored in RSIData.
- Updates: recalculated when each new candle closes for the active timeframe.
OHLC values
OHLC updates as you move across the Original chart. It reflects the candle at the crosshair position, not always the latest live candle.
Use OHLC to verify wick rejections, breakout closes, and volume bar alignment with pattern lines.
Historical insight
Historical insight compares the selected live setup to our offline research for the same pattern type and timeframe. It is educational context — not a prediction or trade signal. Live “breakout direction” on the chart is price-vs-levels bias and is not these rates.
Pattern outlook uses exclusive breakout / breakdown / forming rates from the matched 3-way (else 2-way) category cell for this live setup — the same cell as the typical-travel bands. The bullets describe that cell versus the pattern×timeframe headline.
Coin detail does not load live/geometry feature tables. Category-tree cells stay in the research JSON and drive outlook, matches, and typical-travel bands.
The matched 3-way or 2-way cell sits beside the outlook rates. Other matches (extra 2-ways and 1-way bins used to fill up to five) sit below with the same exclusive breakout / breakdown / forming bars. If fewer than five exist, all of them are shown. Headline rates still come from the primary 3-way (else 2-way) matched category. A feature needs a combined breakout/breakdown rate spread of at least 25 to enter the tree; weaker 3rd features are dropped so some pattern×timeframes are 2-way only. If that 3-way cell is empty, the next 2-way pair is used. Thin cells are kept (no n drop). RSI, Fuel, pattern scale, and category-move bands always use the last closed candle. They rematch when a new candle closes, not on forming-bar ticks. Trendline patterns (triangles, wedges, channels, flags, pennants) use two research versions: while the setup is still forming, rates and typical travel come from forming snapshots only (pattern-end terminal candles are excluded). After a known breakout (first breached candle already above the breakout line), Average up / Average down come from a breakout-first terminal category tree — historical pattern-end snapshots grouped by which trendline closed past by 0.5× NATR first (the same threshold that flags a live recent breakout/breakdown), then scored on their real 1.0× NATR forward outcome. After a known breakdown (first breached candle already below the breakdown line), the mirrored breakdown-first terminal tree is used. If the side is unknown, the mixed pattern-end terminal snapshot is used.
“Full study” opens the public research page for the full methodology and charts.
- Pattern outlook = exclusive rates of the primary 3-way (else 2-way) category cell. The list is up to five matches: 3-way, then 2-way, then 1-way to fill. Other matches show the same exclusive rates as the primary card.
- Baseline (for pp) = the same snapshot universe as the matched cell: forming snapshots while Trendline Forming = Yes; breakout-first or breakdown-first pattern-end terminal after a known side; mixed terminal if the side is unknown; last snapshot for non-trendline types.
- Breakout / Breakdown / Forming or Sideways = exclusive next-20-candle first-hit classes vs 1.0× NATR. The third class is Forming while a trendline is still forming (stayed between the trendlines). Once resolved — after a known breakout/breakdown, or for non-trendline types (head & shoulders, double top, cup & handle, etc.) — the third class is the real Sideways rate (tagged a line, no 1× NATR follow-through). After a known breakout/breakdown, all three rates are measured within the cohort of historical patterns whose first breach (0.5× NATR) was the same side — Breakout is the real share of that cohort that went on to clear 1.0× NATR, not assumed from the side label. An i-icon appears only on a Forming panel if those three rates still sum to less than 100%.
- ×NATR under each class = median wick max-up / max-down in that class, divided by row NATR.
- pp = percentage points vs the pattern×timeframe baseline rates.
- RSI / Fuel / scale / BTC enter the category tree from the last closed candle and rematch when a new candle closes. Formation geometry is last-snapshot. Trendline Forming = Yes → forming snapshots only (exclude terminal). Trendline Forming = No with a known breakout or breakdown → that side's first-breach (0.5× NATR) pattern-end terminal tree only. Unknown side → mixed terminal.
- Matched category = 3-way then 2-way tree; feature spread score must be ≥ 25; no cell n filter. Chart TP/SL band = category median 20-candle wick travel scaled by pattern NATR. Exclusive first-hit gates stay in the research metrics (1.0× NATR) and are not drawn on the chart. Avg up / Avg down bands stay on last-closed close until the next candle closes.
- n = sample size in that bin or category cell.