ROC Advanced

Features

• ROC core: percentage rate of change → averaged main line → signal line via the shared smoothing kernel
• Zero-centered scale: 0 midline with configurable positive/negative stretch defaults at ±5%
• Signals on raw percentage values: crossover detection uses ROC and its signal directly
• 4-state histogram: rising/falling × positive/negative drives four opacity states
• Gradient line: color.from_gradient() from bull teal to bear pink based on positive/negative stretch
• Shadow fills: gradient shadows between ROC and zero
• Signal markers: configurable stretch-zone filter for ROC/signal crosses; optional zero-cross dots
• Stall/absorption layer: central to this variant; flags bars where ROC changes sharply while price barely moves
• Trend context line: a slow ROC plotted faint behind the fast line; a cross against its side of zero is marked counter-trend
• Divergence wedge: fills the area between fast ROC and the trend context only when they move in opposite directions
• Sentiment Bar: optional live label at the panel’s right edge — a signed ±100 score for how far ROC sits inside its own stretch zone, plus a mini bar
• Signal Quality: optional 0-100 score next to each Bull/Bear Stretch marker — stretch-zone depth (50%) + context agreement (25%) + stall-free (25%)
• Alert conditions: bull/bear cross from stretch zone; zero-cross up/down

Scale

This implementation uses classic percentage ROC:

ROC = 100 * (source - source[length]) / source[length]

The default stretch levels are ±5%, intentionally configurable because raw ROC magnitude depends heavily on instrument, timeframe, and volatility. The panel has no fixed outer scale boundary — it autoscales to the actual ROC range so small-magnitude instruments aren’t compressed near zero.

Behavior

ROC is best read as momentum acceleration/deceleration. A large ROC move on flat ATR-normalized price can be a useful stall/absorption clue, while a slow ROC context above or below zero gives the fast cross a directional backdrop.