Taiwan Ownership-Flow Dashboard — three sources side by side, never summed
Comparable products blend institutional flows, holder concentration and futures open interest into a single “chip score” or “main force direction”. The problem is not accuracy but definability: lots, percentages and contracts are three different units; daily and weekly are two different frequencies; single-stock and market-wide are two different scopes. Blending them requires three invented conversion coefficients. This page answers one question instead: as of the query date, where does each source stand and what does it actually say?
Different frequencies, so nothing is force-aligned
Holder brackets are weekly and become available several days after their data date; cash-market institutional flows and futures open interest are daily. The latest available date is usually not the same day across the three.
Each source returns its data date, availability date and lag against the query date, and an alignment block states plainly which day each figure stands on. Upsampling weekly data to daily would manufacture observations that did not exist that day.
Scope labels: the futures column is not about this stock
Futures open interest is market-wide; cash-market flows and holder brackets are single-stock. Every source carries an explicit scope field and a warning, because a market-level position must not be read as this stock's main force direction — the most common category error in flow analysis.
Why there is no chip score and no direction call
The response contains no score, rank, signal or target-price field, and ships the reason for refusing to synthesise.
Two clarifications matter: rising holder concentration does not imply a rising price, and a larger foreign net-short futures position does not imply a falling market — institutional futures positions are frequently the hedge leg of a cash position.