The flow graph shows liquidity buckets and asset classes as nodes, with edges representing the relationships between them. Every element on it is a measurement, and knowing which measurement each element encodes is what makes it readable rather than decorative.
| Element | Encodes |
|---|---|
| Node value (−100 to +100) | Net directional pressure from that node’s weighted factors |
| Node border intensity | Magnitude of that pressure |
| Edge thickness | Measured strength of the relationship in the current window |
| Edge opacity | Confidence, derived from how tightly the two have co-moved |
| Edge direction and colour | Whether capital pressure is building or retreating |
| Observed lag | The lead/lag in days measured from the data, not assumed |
Node pressure is rescaled by the weight of the factors actually observed. If a node expects six inputs and only four are available, the reading is computed from those four and rescaled, rather than treating the two missing series as zeros. The data coverage figure beside the graph tells you how much of the expected input set was present.
The edge from equities to Bitcoin shows low thickness and low opacity. That means the two have not co-moved tightly in the current window and the confidence in the relationship is low. Concluding "equities are up so Bitcoin will follow" would be reading a relationship the graph is explicitly telling you is currently weak.
Common belief
"The graph shows what will happen next."
What is actually true
It shows measured pressure and measured relationships in the current window. Pressure building at a node is a description of conditions, not a forecast of price. The WHY panel beside it exists to make the drivers inspectable so you can disagree with the reading on specific grounds.