A flow of trades and a stock of obligations
Volume can be shown in contracts, underlying units or money. Trade count is another measure: one trade for 100 contracts differs from 100 trades for one contract. Before comparing sources, identify unit, venue coverage, session and report status. A large number without these definitions cannot establish even a change in activity.
Three operations, three outcomes
Suppose open interest in a fictional future is 100 before trading. Two new participants open opposite positions in ten contracts: volume rises by 10 and open interest becomes 110. Two existing participants then close seven: volume rises by another 7 and open interest falls to 103.
A prior long holder transfers five contracts to a new buyer: one participant closes and another opens. Volume rises by 5 while open interest remains 103. Total daily volume is 22 and the change in open interest is +3. This is fictional contract bookkeeping, not a description of actual trading. A large turnover need not create a comparable increase in outstanding positions.
Maturities and seasonal comparison
Near expiry, participants may close one maturity and open the next. Looking only at the nearest contract can make this look like disappearing interest in the whole market. Check several maturities and aggregation rules. Simple summation also does not explain economic exposure: a spread position may have opposite legs at different maturities.
Do not confuse the observation date for volume with the publication date for open interest. A report may contain preliminary and final data or different update timing. For a research test, use the version available at the time under study. A later corrected table is suitable for final facts but may not reproduce the information available then.
What OBV accumulates
In the educational OBV rule, add the period’s volume when closing price rises, subtract it when price falls and leave the value unchanged when the close is equal. The starting level is arbitrary. With fictional closes 100, 102, 101, 101 and subsequent volumes 10, 8 and 6, starting at zero gives 10, 2 and 2. The last volume changes nothing because the close did not change.
This rule assigns all period volume to one sign of the closing-price change. It does not reconstruct intraperiod trades, measure net money inflow or reveal hidden intent. A large positive number depends on the starting point. A claim that OBV is predictive requires an independent test, not merely matching rising lines. Lo and co-authors discuss the broader formalization problem.[2]
What to check before interpreting
Match contract specification, volume units, open-interest date, venue coverage and maturity roll. Check corrections and missing values separately. The [futures and options guide](/guides/futures-and-options/) explains obligations, while [calendar spreads](/guides/calendar-spreads/) explains two maturities. The result should be a correct activity-and-obligations table; future price direction does not follow automatically.
What to ask the data provider
Ask whether the table combines futures and options, which maturities are included and whether volume and open interest refer to the same date. Preserve the original row before transformations. If units or calendar are unknown, fix the data description before interpretation. Keep the [open-interest glossary entry](/glossary/otkrytyy-interes/) beside the table so a stock of obligations is not turned into supposed net cash inflow.
Sources and limits of review
[1] [CFTC Futures Glossary](https://www.cftc.gov/LearnAndProtect/AdvisoriesAndArticles/CFTCGlossary/index.htm): contracts during a period and open contracts; one side is counted for open interest. [2] [Foundations of Technical Analysis](https://www.nber.org/papers/w7613): operationalization of chart patterns and conditional historical informativeness, not proof of profit. The original CME materials could not be obtained in full. The definitions use the complete primary substitutes listed above; only the named concepts and educational arithmetic were checked. Trading interpretations and predictive promises were not validated.
Volume beside a pattern: test a hypothesis, not the future
Consider a breakout from a predefined 100–105 range. Ten prior sessions had average volume of 1,000 shares and the close at 106 had volume 1,800. The ratio 1.8 describes higher activity that day. It does not measure the probability of continuation: the same volume is compatible with exits by old holders, new offsetting positions and a later return to the range.
To make “volume confirms the pattern” testable, predefine range, breakout criterion, averaging window, comparable sessions, outcome horizon and cost treatment. Compare subsequent frequencies with controls, including failed breakouts. Lo and co-authors study formalized chart features, not guaranteed profit from a picture.[22](#source-rm-technical-lo) If the range and convenient spike were chosen after seeing the outcome, the illustration is not an independent test.
ETF: secondary turnover is not creation of units or a NAV change
State Street’s primary explanation says new ETF units are created in large blocks through authorized participants in exchange for a specified basket; redemption reverses the process. Ordinary holder trades occur on the secondary market. Unit liquidity also relates to liquidity of basket components, so one past exchange turnover does not exhaust the assessment.[21](#source-mk-ssga-etf) This is a foreign mechanism, not a universal right for a Russian fund holder to exchange a unit for assets.
In a fictional fund with NAV 100, the same 1,000 units are resold ten times in the morning: volume is 10,000, but issued units and NAV do not increase arithmetically. In another case, underlying assets are revalued by 2% with no secondary trades at that moment; NAV becomes 102 without new exchange turnover. Neither example establishes an exact relation between unit price and number of trades.
If the basket is worth 100 per unit and units sell at 101, creation and sale may offer a gross difference of 1 before costs. With fictional aggregate costs of 0.60, the remainder is 0.40. This supports an incentive, not guaranteed arbitrage: block size, basket, timing, financing and market risk matter.