Order Book Imbalance: How to Read It and Trade It
Order book imbalance compares the size of resting buy orders (bids) with resting sell orders (asks) near the current price. If bids clearly outweigh asks, buyers are showing more size at the levels that matter, and price has a slightly higher chance of ticking up. If asks dominate, the reverse. It is a measure of pressure, not a prediction, and on its own it is easy to misread.
This guide covers the formula, the difference between resting imbalance and traded volume imbalance, what the research says, and how to use it on crypto, forex and stock charts without fooling yourself.
TL;DR
Order book imbalance measures how much size rests on the bid versus the ask. A common version is (bids − asks) / (bids + asks), which runs from −1 to +1.
It is not the same as volume imbalance, which shows where aggressive market orders have already traded.
Academic work on NYSE data links order flow imbalance at the best bid and ask to short-term price changes, but that is a short-horizon, high-frequency finding, not a swing-trade signal.
Resting orders can be pulled before they trade, so treat a one-off lopsided book as noise until it repeats or gets confirmed by executed volume.
Test any imbalance rule on paper before you risk money, and size positions for the case where it is wrong.
Table of Contents
What order book imbalance actually measures
An order book is a live list of limit orders. Bids are buyers waiting at or below the market. Asks are sellers waiting at or above it. The imbalance asks one question: of the size resting near the price, how much is on each side?
The simplest statistic, as laid out in this walkthrough of order book imbalance, is:
OI = (B − A) / (B + A)
where B is bid volume and A is ask volume over the window you choose. The result sits between −1 and +1. A reading of +0.60 means substantially more displayed size on the bid than on the ask.
The same source separates three variants worth knowing:
Level-1 imbalance: only the best bid and best ask. It shows the most immediate pressure.
Multi-level imbalance: the sum across the first k price levels. It shows broader positioning.
Depth-weighted imbalance: gives more weight to orders closer to the midpoint, because those are the ones most likely to trade.
A quick worked example. Say the top five levels of a BTC/USDT book hold 120 BTC of bids and 40 BTC of asks. OI = (120 − 40) / (120 + 40) = 0.50. Buyers show three times the size. If the book later shows 60 and 60, OI is 0 and the edge, if there was one, has gone.
The number depends entirely on your choices: how many levels, whether you weight by distance, and how often you sample. Two traders can both say “the book is bid-heavy” and mean different things.
Order book imbalance vs. volume imbalance
These two get confused constantly, and they answer different questions.
Order book imbalance | Volume imbalance | |
What it counts | Resting limit orders (intent) | Executed aggressive market orders (action) |
Where you see it | Depth of market (DOM), order book heatmap | Footprint charts, volume at price |
Can it vanish? | Yes, orders can be cancelled | No, the trades happened |
Time horizon | Seconds to minutes | Bars, sessions, levels on a chart |
As one knowledge-base explanation of volume imbalance puts it, resting imbalance shows current liquidity pressure at price levels, while volume imbalance shows where market orders have already executed successfully. Resting size is a statement of intent. Executed volume is evidence.
If you want a refresher on why that distinction exists, our beginner’s guide to market vs limit orders covers who supplies liquidity and who takes it.
The diagonal comparison most guides skip
When you do read volume imbalance on a footprint chart, there is a technical detail that changes the result. According to this explanation of footprint imbalances, the comparison should be diagonal: market buys at price P are compared with market sells at price P−1. An aggressive buyer lifts the ask, and to push through it must consume resting liquidity on the level below. Comparing buys and sells on the same row ignores that one-tick gap and can flag imbalances that are not really there.
If your charting tool offers the choice, use the diagonal method, and check which one a given indicator uses before you trust its markers.
What the research says
The strongest evidence for imbalance is from market microstructure. A study in the Journal of Financial Econometrics, Price Impact of Order Book Events, used NYSE Trades and Quotes data for fifty U.S. stocks. It found that over short intervals, price changes are mainly driven by order flow imbalance: the difference between supply and demand at the best bid and ask. The relationship was roughly linear, with a slope inversely related to market depth. Thin books move more for the same imbalance. Deep books absorb it.
Two cautions before you take that to your trading screen.
The horizon is very short. The finding concerns short intervals, the world of market makers and high-frequency systems. It does not say a bid-heavy book on a 4-hour chart predicts a rally.
Net of costs, it is hard. A statistical relationship in price changes is not the same as a tradeable edge after fees, spread and slippage. Our slippage analysis playbook shows how to measure what execution really costs you.
So the honest reading is this: imbalance is a real, measured driver of tiny price moves, and a useful input for understanding who is pressing. It is not a standalone system.
A note on “order imbalance” in US stock markets
If you trade stocks, you may see the same words used differently. On the NYSE, “imbalance” is a defined term in the auction process. Under NYSE Rule 7.35, auction imbalance is the net volume of buy or sell shares that cannot be paired at the indicative match price at a given moment, including the non-displayed quantity of eligible Reserve Orders. Market Imbalance refers specifically to the remaining market orders that cannot be matched.
In practice that means the opening and closing auctions publish imbalance information that tells you which side is unfilled before the auction prints. That is a different object from a continuous order book ratio. Same word, different mechanics. When you read a stock-market imbalance headline, check which one it is.
Crypto exchanges run continuous order books around the clock with no official auction, so the ratio and footprint methods above are what apply there.
How to read an imbalance without being fooled
A lopsided book looks informative. Often it is not. These are the checks that separate signal from noise.
1. Ask whether the size can disappear
Resting orders cost nothing to cancel. A large bid that sits several levels below the market can be pulled as price approaches. This is why a single snapshot says little. Look at whether the size stays, gets filled, or evaporates when price nears it.
2. Look for repetition, not a single print
An imbalance that is never revisited is usually noise. The same knowledge-base source notes that professionals look for clustered imbalances that repeatedly show one-sided dominance, because those levels often act as real support or resistance when price comes back. One stray marker is a curiosity. Three at the same zone is a level.
3. Confirm with executed volume
If the book shows heavy bids and sellers are still hitting the bid and getting filled with price falling, the bids are being absorbed and may fail. If price holds while sell market orders pile in, that is absorption in your favor. Pairing resting imbalance with traded volume imbalance is far stronger than either alone.
4. Account for depth
The same imbalance means different things in a deep book and a thin one. A 3-to-1 ratio on a thinly traded altcoin pair may be a few hundred dollars of size. On a major pair it is serious. Always read the ratio next to the absolute size.
5. Place it in context
Imbalance is a local signal. Trend direction, market structure and volatility decide whether it matters. A bid-heavy book inside a strong downtrend often means buyers are being run over, not that a bottom is in. This is why we pair order flow reading with structure concepts such as CHoCH and fair value gaps.
Using imbalance in a trading plan
Here is a simple, testable structure. It is an illustration of how to build a rule, not a recommendation of specific numbers.
Define the measurement. For example, depth-weighted OI over the first ten levels, sampled at a fixed interval.
Define the trigger. For example, OI above a threshold you chose in advance for a set number of consecutive samples.
Add confirmation. Require diagonal volume imbalance in the same direction at the same price zone.
Add a location filter. Only act near a level you marked beforehand, such as an order block or prior swing.
Define the invalidation. Decide where the idea is wrong, and size the position so that loss is acceptable.
Record every instance. Log wins and losses, including the ones you skipped.
The last step is where most traders cut corners. Without a record, you cannot tell a real edge from a lucky week. Judge the rule on sample size and expectancy, not on how many memorable trades it produced.
Execution matters more than the signal
Imbalance signals tend to be short-lived, so execution quality can decide the outcome. A market order into a thin book pays the spread and slips. A limit order may never fill. Our piece on order fill analysis explains how to measure fill probability so you can see whether a signal you like is actually capturable.
Where imbalance strategies break
Spoofed or transient liquidity. Large orders placed to be cancelled distort the ratio. This is another reason to weight executed volume above resting size.
Fragmentation. Crypto liquidity is spread across many venues. One exchange’s book can be lopsided while the aggregate is balanced.
News and volatility spikes. In fast markets books thin out and ratios swing wildly. A rule tuned for quiet conditions can fail badly. Our guide to the volatility breakout strategy discusses how conditions change what works.
Overfitting. Tuning the number of levels, thresholds and windows until the backtest looks good is easy. Out-of-sample results usually look worse. A systematic approach means fixing the rules before you look at results.
Retail data limits. Most mobile and retail platforms do not give you the full, timestamped order book that the academic studies used.
Our perspective
We build Discipline AI for traders who want evidence, not stories, so here is how we think about imbalance.
Imbalance is a useful clue about conditions, and a poor reason to enter a trade by itself. In our own research approach, we look at market structure, liquidity events, trend direction, volatility and market regime together, and we judge a setup by what the historical outcomes say: win rate, risk-adjusted expectancy, reward-to-risk and, above all, sample size. A pattern that worked eleven times tells you very little. A pattern that has been measured across many instances, in the regimes where it applies, tells you something.
We also think the most underrated skill is standing aside. A lopsided book in a hostile regime is a reason for caution, not action. Much of trading discipline is declining marginal setups, and that is why our platform treats stand-aside conditions as a feature rather than a failure to find a trade.
And no indicator, ours included, guarantees profit. The honest goal is to make decisions that hold up when you review them later.
— Tony
Practice it before you risk it
If you want to test an imbalance idea, the cheapest way is to practice on paper and review the results honestly. Discipline AI is a mobile app for iOS and Android that combines AI-generated trade setups with confidence scoring, risk tools such as position sizing and leverage calculation, paper trading with P&L tracking, automated journaling and AI Trade Autopsies that review completed trades. Our Sentinel research system evaluates market structure, liquidity events, volatility and regimes, so you can see whether conditions favor your idea before you act.
Pro is $8.99 per month, $79.99 per year, or $199.99 one-off. It does not promise profitable trades. It gives you a structured way to find out whether your approach has an edge.
FAQ
Is order book imbalance a good indicator?
It is a real but short-lived one. Research on NYSE data ties order flow imbalance to price changes over short intervals, but that does not make it a reliable standalone signal on longer timeframes. It works best as one input alongside executed volume, structure and risk limits.
What is the difference between order book imbalance and volume imbalance?
Order book imbalance compares resting limit orders on the bid and ask, so it reflects intent that can be cancelled. Volume imbalance compares aggressive market orders that have already executed, usually on a footprint chart. The second is evidence of what happened. The first is a hint of what might.
How do you calculate order book imbalance?
A common formula is (bid volume − ask volume) / (bid volume + ask volume), giving a value from −1 to +1. You choose how many price levels to include and whether to weight closer levels more heavily. Document those choices, because they change the result.
Does order imbalance mean the same thing on the NYSE?
Not quite. On the NYSE it is a defined auction term: the net buy or sell shares that cannot be matched at the indicative price in an opening or closing auction. A continuous order book ratio in crypto is a different measurement.
Sources
Recommended
Comments