💡

TL;DR: The Institutional Ranking Board (/vip/institution) ranks all ~2,000 listed Taiwan stocks by net buy/sell from foreign investors, investment trusts, and dealers. ETFs are excluded by default to avoid market-maker distortion. Expand the full top 100 and resort by any column. The fastest way to "follow smart money for clues."

Concepts

Why a Whole-Market Ranking Instead of Per-Stock Flows?

Per-stock institutional data tells you "how much institutions bought of this stock today." But it can't tell you "which stocks institutions bought most today." The latter is what unlocks new ideas:

A stock outside your watchlist will never appear on your radar otherwise
Stocks bought heavily on consecutive days often mark trend starts
The board reveals where capital is flowing at a sector level (e.g. all top 20 are financials on a given day)

Why Exclude ETFs by Default?

ETFs (especially large ones like 0050, 00878, 00919) generate massive daily passive trades by their market makers. The dollar volumes routinely dwarf any individual stock. Without exclusion, the top 5 net-buy positions would almost always be ETFs — not useful for finding stock-level opportunities.

You can still toggle "Include ETFs" to view ETF flows themselves (especially active ETFs like 00981A).

How the Ranking Is Calculated

After market close (around 17:00 local), TWSE and TPEX publish the day's institutional data. We ingest it and compute:

Foreign net: Foreign (incl. foreign dealer arms) total buy − total sell
Trust net: Investment trust total buy − total sell
Dealer net: Dealer (proprietary) total buy − total sell
Big-3 combined: Sum of all three

Default sort is by combined net descending; you can click any column header to resort.

Hands-on: Reading It on CTSstock

Visit /vip/institution (VIP feature). Three boards are shown:

Net-buy leaderboard (combined net buyers)
Net-sell leaderboard (combined net sellers)
Single-institution boards (foreign / trust / dealer separately, switchable)

4 Usage Patterns

Pattern 1: Discover new ideas (10s)

Open the net-buy board. Scan the top 10 for stocks not on your watchlist. Click an unfamiliar one and spend 60 seconds — sector, market cap, recent price action. If preliminarily acceptable, add to watchlist for observation.

Pattern 2: Look for continuity (1 min)

Single-day rankings are noisy. A stock with a 500M NTD net buy today might mean nothing; appearing in the top 30 for 5 consecutive days is a real trend. Check daily and note repeats — only research stocks that stay on the board for 3-5 days.

Pattern 3: Cross-reference (3 min)

Pair the buy and sell boards:

Foreign + trust both buying → strongest consensus
Foreign buying but trust selling → divergence, dig deeper
Dealer-driven names → likely short-term theme plays, follow with care

Pattern 4: Expand 100 rows for filtering

Hit "Show all" for the full 100. Then:

Sort by "Dealer net buy" → catches small-cap themes that foreigners ignore
Sort by "Volume" → spots low-volume names suddenly seeing institutional entry
Switch to OTC (TPEX) — institutional moves on OTC stocks often precede those on listed (TWSE)

FAQ

Q: Does institutional buying always lead to gains?

No. There are several motives:

Bullish on fundamentals — higher win rate, but takes time to play out
Hedging / arbitrage — e.g. foreign desks buying spot to hedge index futures, no directional view
Passive index tracking — MSCI/FTSE rebalances trigger mechanical buys

To judge whether institutions "really like" a stock, look at continuity + synchronization — multi-day buying from multiple institutions is the strong signal.

Q: How is the ETF exclusion done?

Filtered by the index field in the basic_info table — anything containing "ETF" is dropped. So ETNs, warrants, and other special instruments are also excluded. Toggle the switch to include them.

Q: Are 100 rows enough? Should I see 200?

100 is plenty in practice. Institutional capital concentrates in the top 50; rows 50–100 already see net flows under 10M NTD with limited reference value. If you genuinely need beyond row 100, full CSV export is on the roadmap.

Q: Are dealer rankings worth watching?

Yes, but carefully. Dealer activity has two parts:

Market-making — hedging, no directional meaning
Proprietary — corporate funds going long or short with conviction, this part has signal value

Public data doesn't distinguish them, so treat the dealer board as a supplementary signal — only when a stock appears on both foreign and dealer boards does it become a strong combined signal.