Volume counts shares traded during an interval. Both a buyer and a seller participate in each trade. Higher total volume alone does not identify institutional net buying. This site contains neither participant-level orders nor an accumulation detector.
Separate multiples from increases
Hypothetical daily volume of 1.2 million versus a 50-day average of 800,000 is 1.5 times average, or a 50% increase. “150% of average” differs from “a 150% increase.” The screener JSON stores average volume in thousands, rounded down: 800 means 800,000 shares, not that day's volume.
Check liquidity separately from qualification
On 2026-10-02, 282 of 523 U.S. passers met a 500,000-share average-volume threshold (53.92%). Korea had 21 of 90 (23.33%). This is a share-count filter, not a comparison normalized for price, currency or dollar turnover. Average volume does not disclose the spread or the capacity for a particular order.
Observation is not buyer identification
Up-day and down-day volume summarize observations. Index changes, announcements, crossing trades and share-count changes can produce similar patterns. Institutional holdings require separate filings. U.S. filings can be researched through SEC EDGAR; align the filing date with the price observation date before interpreting either.
Check anomalies
A zero caused by a trading halt differs from a missing observation. Verify split treatment, whether the event day enters the comparison average, and whether daily and weekly periods were mixed. Provider volume is a summary, not evidence identifying the participants.
Data and corrections
See the methodology for calculation rules and the 2026-10-02 signal-disagreement study for the frozen sample. Report a date and URL through contact when you find an error.