How to read price and open interest together
Measured in contracts rather than dollars, open interest added no significant 24-hour signal beyond the price move itself. Tested across five Binance markets.
The direct answer
Open interest tells you whether the quantity of outstanding derivative contracts expanded or contracted. It does not tell you whether longs or shorts are "in control," who initiated the trades, or whether the new positions are well informed.
Price supplies direction. OI supplies the change in outstanding exposure. Read together, they describe whether a move happened while contracts were being added or removed. The familiar participant stories attached to those quadrants remain hypotheses until other data supports them.
In our five-market test, price down + OI down finished above each market's overall 24-hour hit-rate baseline, while price up + OI up finished below it in four of the five markets and level with it in SOL. But almost all of that ordering came from the price move itself. Once up hours were compared with other up hours and down hours with other down hours, contract OI added no conventionally significant directional information.
That is still useful. It tells you to use OI for describing leverage expansion and contraction, not as a standalone forecast.
How open interest changes
Every open futures contract has a long side and a short side. One cannot exist without the other. OI counts the outstanding contract exposure, not a surplus of longs over shorts.
When a trade occurs, the opening and closing status of both sides determines what happens to OI:
- A new long matched with a new short increases open interest.
- An existing long closing against an existing short closing decreases open interest.
- A participant opening against another participant closing leaves total open interest unchanged.
This is why rising OI does not identify the aggressor. New contracts can be created while aggressive buyers lift offers, while aggressive sellers hit bids, or while activity alternates between both.
Falling OI has the same limit. It says contracts disappeared on net. It does not tell you whether they closed voluntarily, through stop orders or through liquidation.
Contract OI and dollar OI are not the same
An OI chart can be denominated in contracts or base-asset quantity, or it can be marked into dollars.
Dollar open interest changes for two reasons:
dollar OI = base-asset OI x current price
If BTC price rises while the number of open BTC contracts stays flat, dollar OI rises mechanically. If price falls, dollar OI falls. That makes dollar OI useful for estimating current notional exposure, but dangerous in a study asking whether OI changed independently of price.
That is easy to state and easy to underestimate, so here is the size of it. On the 41,702 BTC hours used below, the hourly change in dollar OI correlates with that hour's price return at 0.610. The same change measured in contracts correlates at 0.096. Sorting hours by dollar OI is therefore closer to sorting them by price than most chart readers would guess.
The effect on the quadrants is direct. Among hours where price rose, 73.6% are classified as rising OI when dollar denomination is used, against 53.5% when contracts are used. Nearly three quarters of up hours land in the price up + OI up box automatically, whether or not a single contract was created.
quantumflow.techThe historical test below uses base-asset open interest. It therefore classifies whether outstanding contract quantity expanded or contracted without marking the same positions up and down with the price move being studied.
One rule, two contract designs
The relationship above holds for linear contracts like Binance's USDT-margined perpetuals, where the position is denominated in the base asset and its dollar value is that quantity marked at the current price. Every market in this study is one of those.
Inverse, coin-margined contracts work the other way. A contract there represents a fixed dollar face value, so its dollar open interest does not rise merely because price rose. The base-asset value of the same position moves inversely with price instead.
So the rule is not "always prefer contracts to dollars." It is to use whichever unit the contract holds fixed:
- Linear, base-denominated products. Read base-asset or contract quantity. Dollar OI carries the price move inside it.
- Inverse, dollar-denominated products. Read contract count or the fixed dollar face amount. The base-asset figure is the one contaminated by price.
- Across venues. Contract counts are not comparable until each product's multiplier is normalized. One contract may represent a dollar, a hundred dollars, or a whole coin.
When reading a chart, check the unit before interpreting the quadrant. A price rise with rising dollar OI is not proof that contracts were added.
The four familiar quadrants
| price | open interest | what can be observed | common interpretation |
|---|---|---|---|
| up | up | price rose while contracts were added | trend participation |
| up | down | price rose while contracts were closed | short covering |
| down | up | price fell while contracts were added | new bearish positioning |
| down | down | price fell while contracts were closed | deleveraging or exhaustion |
The observable column is mechanical. The final column is a hypothesis.
Price up, OI up
Price rose while more contracts remained open. That is consistent with new participation, but every new contract still contains a long and a short. The quadrant does not tell you whether buyers chased, sellers supplied them or the positions are likely to persist.
Price up, OI down
Price rose while contracts closed on net. Short covering is one possible cause, but OI does not identify which side initiated the closing flow. Longs taking profits and a broader reduction in exposure can occur in the same window.
Price down, OI up
Price fell while outstanding contracts expanded. New bearish positioning is a common reading, but the contract creation still requires both sides. The quadrant cannot separate aggressive shorts from passive longs absorbing them.
Price down, OI down
Price fell while contracts disappeared. Liquidations, stops and voluntary exits can all contribute. Calling the state capitulation requires evidence that the forced or vulnerable exposure was actually removed, not only a falling OI line.
The largest one-hour fall in BTC contract open interest in the whole sample is below, picked by that rule rather than by eye. The tests further down measure whether hours like it were followed by anything systematic.
quantumflow.techA practical way to use the quadrants
The study below does not validate a quadrant trading strategy. After controlling for the sign of the price move, neither OI hit-rate difference was conventionally significant. These steps use OI to organize evidence rather than turn it into a forecast by itself.
1. Check the OI unit. Read whichever unit the contract holds fixed: base quantity on a linear product, contract count on an inverse one. Use dollar OI when asking how much current notional is exposed, and remember that on a linear product price itself changes it.
2. State only the mechanical observation. Say "price rose while contracts expanded," not "healthy longs entered." Keep the participant story separate.
3. Use liquidations to explain contractions. A large OI decline accompanied by forced-close prints differs from one with little reported liquidation. The remainder can still include stops and voluntary exits.
4. Use the tape to identify aggression. Spot and perp CVD can show which side crossed the spread. They still do not identify the passive counterparty or prove intent.
5. Put funding against its own baseline. Unusual carry can describe the cost and market state around the open contracts. Positive funding alone does not prove the growing side is crowded.
6. Let price levels define the decision. OI can tell you exposure expanded into a breakout or contracted during a rejection. The level holding or failing is the observable invalidation; the OI quadrant is supporting context.
How we tested the framework
We used Binance perpetuals from 2022-01-01 through 2026-10-15. Each clock hour uses the first and last one-minute candle prices and the final base-asset open-interest reading. The OI change is calculated only when the preceding observation is the immediately preceding clock hour. The outcome price must exist exactly 24 clock hours later.
Those adjacency rules leave 41,702 usable BTC hours. They also explain why the price-only and quadrant tables below have identical sample counts.
BTC perpetual:
| quadrant | hours | avg OI change | higher 24h later | mean 24h return |
|---|---|---|---|---|
| price up + OI up | 11,273 | +0.44% | 50.31% | +0.073% |
| price up + OI down | 9,781 | -0.42% | 49.92% | +0.046% |
| price down + OI up | 10,418 | +0.39% | 52.02% | +0.078% |
| price down + OI down | 10,230 | -0.47% | 52.33% | +0.049% |
Price up with rising OI is below the BTC baseline, and price down with falling OI is above it. Notice that the four cells hold roughly similar numbers of hours. That is what you expect when the open-interest classification is not being driven by the same price move that defines the row.
quantumflow.techWhat repeated across five markets
| market | up + OI up | up + OI down | down + OI up | down + OI down | baseline |
|---|---|---|---|---|---|
| BTC | 50.3% | 49.9% | 52.0% | 52.3% | 51.1% |
| ETH | 50.0% | 49.6% | 51.2% | 51.9% | 50.7% |
| SOL | 49.2% | 48.6% | 49.5% | 49.2% | 49.1% |
| XRP | 47.4% | 48.2% | 49.2% | 49.2% | 48.5% |
| DOGE | 47.4% | 49.3% | 49.4% | 49.2% | 48.8% |
Two narrow statements describe the five rows:
- Price down with OI down is above that market's overall hit-rate baseline in all five.
- Price up with OI up is below that market's baseline in BTC, ETH, XRP and DOGE, but not in SOL.
SOL's price up with OI up cell is 0.04 percentage points above its baseline, which is why it displays as 49.2% against 49.1% in the table. That gap is far too small to resolve, so SOL is best read as level, not as a reversal of the pattern.
The stronger four-cell ordering does not repeat. DOGE's highest cells are price up with OI down and price down with OI up, while SOL's lowest cell is price up with OI down.
The assets also share crypto market regimes. The repeated direction is context, not five independent confirmations.
quantumflow.techPrice direction explains the pattern
On the identical 41,702-hour BTC sample:
| the hour was | hours | higher 24h later |
|---|---|---|
| up | 21,054 | 50.13% |
| down | 20,648 | 52.17% |
That 2.04-point spread requires no OI data. Down hours were followed by a higher price more often than up hours, consistent with short-horizon price reversal.
Inside each price direction, OI adds no resolved difference:
- After up hours, the hit rate was 49.92% when OI fell and 50.31% when OI rose. Falling minus rising is -0.38 percentage points.
- After down hours, it was 52.33% when OI fell and 52.02% when OI rose. The difference is +0.31 points.
The 24-hour outcomes overlap. Using Newey-West standard errors with 24 hourly lags, the up-hour difference has a 0.75-point standard error and t = -0.51. The down-hour difference has a 0.77-point standard error and t = 0.40.
Mean-return comparisons are equally weak. Falling minus rising OI is -0.027 percentage points after up hours with t = -0.70, and -0.029 points after down hours with t = -0.71.
The data supports a direct conclusion: the apparent 24-hour quadrant ordering comes from the sign of the price move. Base-asset OI did not add a statistically resolved directional effect inside either price direction.
quantumflow.techThe endpoint gap varies by year
The gap between down + OI down and up + OI up remained positive in each calendar slice, but its size changed:
| year | up + OI up | down + OI down | gap |
|---|---|---|---|
| 2022 | 47.0% | 47.1% | +0.1 pp |
| 2023 | 52.4% | 55.0% | +2.6 pp |
| 2024 | 52.5% | 54.7% | +2.2 pp |
| 2025 | 49.5% | 53.1% | +3.7 pp |
| 2026 | 49.9% | 51.8% | +2.0 pp |
The endpoint gap remains descriptive. It compares an up-hour cell with a down-hour cell, so it does not isolate OI from the price-reversal pattern.
Open interest only becomes useful next to what explains it. QuantumFlow puts OI, funding, liquidations and the order book on one chart, so changes in leverage can be read in context.
Common mistakes
- Reading dollar OI as contract creation. Dollar notional moves with price even when base-asset OI is unchanged.
- Saying more longs than shorts entered. Every new contract has both sides.
- Naming the aggressor from OI. OI reports the stock of contracts, not who crossed the spread.
- Calling every OI decline liquidation. Voluntary closing and stops also reduce outstanding contracts.
- Treating the four quadrants as forecasts. In this sample, price direction explained the 24-hour ordering without incremental OI information.
- Comparing raw OI across assets. Contract specifications and asset prices differ; changes should be interpreted within a consistent unit and market.
- Ignoring the horizon. An hourly classification does not describe every position-building sequence that spans several hours.
Limits
- Twenty-four hours is one horizon.
- Clock-hour classification can split a position-building episode across two rows.
- Every market here is a Binance USDT perpetual.
- The tests establish associations, not participant identity or causation.
- Neither hit-rate difference is conventionally significant after accounting for overlapping outcomes.
---
*Written by Dom at QuantumFlow. Published 5 August 2026, updated 16 October 2026. Figures use Binance perpetual base-asset open interest and one-minute candles from 2022-01-01 to 2026-10-15.*