Why footprint and CVD signals fail

Crypto liquidity is fragmented. A Binance spot divergence can be accurate and still fail to matter when the relevant move is being led in perpetual futures or on another venue. First define which market the data represents, then decide whether it is relevant to the instrument being traded.

Cell height, aggregation and imbalance ratio decide what becomes visible. A threshold that marks every bar is no longer selective; a threshold that marks nothing has hidden the evidence. Record the settings with every review and avoid carrying a ratio from one market or timeframe into another without testing it.

Repeated hits at one price can be consistent with an iceberg, but the tape cannot prove that a hidden order was replenished. Heavy balanced volume can be consistent with absorption, but can also result from a queue changing hands. Use “consistent with” when the data cannot distinguish the alternatives.

Exhaustion can appear before another impulsive leg. A stack can be consumed instead of defended. The reading becomes actionable only when price accepts or rejects the level in the direction required by the plan. If confirmation arrives too far from invalidation, passing is a valid result.

A marker is not a stop. Define in advance what price behaviour disproves the specific claim, and size the position so that a normal loss is survivable. Nothing in order flow supplies a guaranteed win rate or target.

• TradingView Volume Delta methodology

• What the data does and does not cover
• Context-first footprint workflow
• Define invalidation before entry