Order Book: How to Read Market Depth and Estimate Slippage

What the Order Book Shows and What It Does Not Show

The order book collects all limit orders for an instrument and groups them by price. Buyers appear on the left or at the bottom, sellers on the right or at the top, and each level shows the total volume. Trading platforms display only a snapshot of this queue around the current price, and the depth of that snapshot depends on the venue and the market data subscription.

You should learn what the order book does not show before learning how to read it. It contains no commitments: a limit order can be canceled with one click and disappear before you have time to react. It does not show hidden volume: iceberg orders and part of the flow from over-the-counter venues do not appear in the standard snapshot. Nor does it show time: the queue represents the current second, not an intention to remain there all day.

This leads to a practical rule. The order book is good at answering “how much will it cost me to enter right now?” and poor at answering “where will the price go?”

How to Read Bid and Ask Orders

The bid is the price and volume of buy orders. The ask is the price and volume of sell orders. The best bid is at the top of the buyer side, while the best ask is at the bottom of the seller side, and they are always separated: if they matched, a trade would already have taken place.

Farther from the best prices are worse levels: buyers are willing to buy at lower prices, while sellers are willing to sell at higher prices. The volumes at these levels are added together to produce the depth on that side. When a trader says, “there are one hundred lots for sale in the order book,” they almost always mean the total across several top levels, not just the best ask.

Another detail that trips up beginners is that a market buy order executes against the ask, not the bid. You buy from someone who is selling. This is why the entry price for a Long position is generally higher than the midpoint shown on the chart.

What Are the Spread and Market Depth?

The spread is the difference between the best ask and the best bid. It shows how much you lose if you enter with a market order and immediately exit again. For a liquid instrument, the spread usually remains within one or two price increments; for a thinly traded instrument, it can be many times wider, and it is easier to check the actual figure in the order book than to take it from an article. The spread changes with liquidity and volatility: it is narrow in a calm market and widens around news events.

Depth is the total volume of orders across several levels on each side. It answers a different question: how far through the price levels an order of your size will reach. A narrow spread with thin depth is deceptive: the first one hundred units may be bought at a good price, while the next five hundred may execute much higher.

These two values must be measured together. The article on the liquidity indicator explains how to view the current liquidity conditions directly on a chart.

123

Spread of one price increment

Spread of dozens of price increments

1. 1.A liquid instrument during the day
2. 2.The same instrument at the open
3. 3.A thinly traded instrument at night

A liquid instrument during the day

The same instrument at the open

A thinly traded instrument at night

Diagram 1. The cost of immediate entry at different spread widths. Source: FINRA Rule 5310, which explicitly lists price, volatility, and relative liquidity among the factors affecting best execution.

How to Trade Using the Order Book

The order book has few practical uses, and all of them concern execution:

• calculate the spread and decide whether entering with a market order is worthwhile at all;
• add up the volume across levels and estimate the average price for your order size;
• compare the total volume on both sides and identify an imbalance;
• choose between a limit order and a market order based on the cost created by that imbalance;
• check whether a large order remains in the book until the next snapshot;
• compare the order queue with the time and sales feed of completed trades.

Everything else commonly presented as an answer to the question of how to trade using the order book comes down to trying to guess the direction from the bars. The rest of the article will show why this is the most expensive part of the method.

Why a Large Order Does Not Guarantee a Price Move

A large volume at a single level looks like a wall from which the price should bounce. Sometimes it does. But a limit order exists only until it is canceled, which makes it a convenient tool for creating an appearance.

In its September 29, 2020 order against JPMorgan Chase, the U.S. Commodity Futures Trading Commission described the mechanics directly:

“other traders often considered information about order book balance when making their trading decisions”

• CFTC, JPMorgan Chase order, 29.09.2020, https://www.cftc.gov/media/4826/enfjpmorganchaseorder092920/download

The same document describes the purpose of such orders as creating a false impression of supply or demand. In other words, enough people watch large orders to make displaying them worthwhile.

The practical conclusion is simple: a wall is confirmed not by its size, but by whether it survives several consecutive snapshots and whether real trades execute against it in the time and sales feed. Similar logic applies to price levels, as explained separately in the article on false breakouts.

How to Calculate Volume Imbalance Across Levels

Imbalance is the normalized difference between the depth on the two sides. Take the five nearest bid levels and add their volumes to obtain the buyer-side depth. Do the same for the ask. Then subtract one from the other and divide by their sum.

The resulting value ranges from minus one to plus one. A positive value indicates a buyer imbalance, a negative value indicates a seller imbalance, and zero indicates equilibrium. Five levels are a compromise: a single level contains too much market noise, while twenty levels already include orders that no one intends to execute.

Research into market microstructure supports the general idea of imbalance, although it uses a different calculation. The study by Cont, Kukanov, and Stoikov measures order flow imbalance at the best bid and ask and shows that short-term price changes are mainly driven by it, while the slope of this relationship is inversely proportional to market depth. Using five levels instead of one is a simplification we adopt to make the picture more stable; it does not come from that study:

“price changes are mainly driven by the order flow imbalance... with a slope inversely proportional to the market depth”

• Cont, Kukanov, Stoikov, arXiv, https://arxiv.org/abs/1011.6402

1. 1

   Add the volumes of five bid levels

   This gives the buyer-side depth

   ↓
2. 2

   Add the volumes of five ask levels

   This gives the seller-side depth

   ↓
3. 3

   Subtract the ask depth from the bid depth

   The sign of the difference shows which side has the imbalance

   ↓
4. 4

   Divide by the sum of the two depths

   The result ranges from minus one to plus one

Diagram 2. The procedure for calculating volume imbalance. Source: our own procedure inspired by the order flow imbalance of Cont, Kukanov, and Stoikov, who calculate it using the best bid and ask.

How to Estimate Slippage Before Placing a Market Order

Slippage is the difference between the price you expected and the price at which the trade was actually executed. An exchange help article defines it as follows:

“Slippage is the difference between a trade's expected or requested price and the price at which the trade is effectively executed.”

• Binance, slippage help article, https://www.binance.com/en/support/faq/what-is-slippage-01f6dd67d54e4dca902914700818e739

The mechanics are clear from the exchange's own figures. An order for 100 coins was filled in parts: 50 at 20 000, 25 at 20 001, and 25 at 20 002. The average price was 20 000.75, meaning slippage was 0.75 per coin relative to the best price of 20 000.

The U.S. Securities and Exchange Commission describes the same mechanics for stocks and provides a separate warning:

“However, the price at which a market order will be executed is not guaranteed.”

• SEC, Trading Basics, https://www.sec.gov/files/trading101basics.pdf

You can calculate this before submitting the order. Take your order size and work through the ask levels from best to worst. At each level, take the available quantity and multiply it by that level's price. The sum of these products divided by the order size is your average price. The difference between this and the starting ask is the expected slippage.

The calculated figure will also be useful later. TradingView's strategy tester has separate slippage and commission fields, and entering your own figure from the order book is more realistic than leaving the default value at zero; otherwise, the test operates in a market without trading costs.

1. Level 1

   All volume at the best ask is taken

   This is the only price you saw on the screen

2. Level 2

   The order moves to a higher price

   The remaining volume looks for the next seller

3. Level 3

   The price moves farther from the starting point

   Each step increases the average price

4. Level 4

   The remainder is filled at the worst price in the set

   On a thin market, the step can be large

5. Result

   The volume-weighted average price is calculated

   The difference from the best ask is the slippage

Diagram 3. How a market order moves through ask levels. Source: the partial execution example in SEC materials and the numerical example in the Binance help article.

How a Limit Order Differs from a Market Order

The trade-off can be expressed in two lines. A market order is almost certain to execute, but its execution price is unknown in advance. A limit order guarantees the price but may not execute at all:

“A limit order is not guaranteed to execute. A limit order can only be filled if the stock's market price reaches the limit price.”

• SEC, Trading Basics, https://www.sec.gov/files/trading101basics.pdf

The U.S. regulator adds an important detail about stop-loss orders: the stop price is a trigger, not an execution price. Once triggered, a stop-loss order becomes a market order and moves through the levels in exactly the same way. The article on stop-loss order types and execution risk explains how this affects the placement of a protective order.

On the Moscow Exchange, this has a direct infrastructural consequence. The exchange limits how far a market order can move from the best price and cancels any unfilled remainder:

“in the event of partial execution, the unfilled remainder is canceled”

• Moscow Exchange, https://www.moex.com/a8519

How to Use the Order Book Together with the Time and Sales Feed

The order book and the time and sales feed answer different questions. The queue contains orders that may disappear a second later. The time and sales feed records trades that have already occurred and cannot be canceled.

A useful combination works as follows. You see a large volume on the seller side. If purchases are simultaneously executing at the ask in the time and sales feed and the order's volume is shrinking, it is genuinely being filled. If the volume remains unchanged and there are no trades, the price has not reached it yet or it is being deliberately held there.

The opposite situation occurs more often than it seems: a large order disappears without a single trade. That is a cancellation, and it tells you exactly nothing about the seller's willingness to trade. The article on the volume indicator separately explains how to read volume flow on a chart.

A Practice Test Using Ten Order Book Snapshots

The following procedure can be repeated in twenty minutes. It does not make money and is not supposed to: its purpose is to show how stable the conclusions commonly drawn by eye actually are.

Choose one liquid instrument and capture ten consecutive order book snapshots at equal intervals, such as once per minute. For each snapshot, record five figures: the best bid, the best ask, the total volume across five bid levels, the total volume across five ask levels, and the imbalance calculated using the formula from the section above. In a separate column, record the largest order and its price.

1. 1

   Choose one liquid instrument

   Switching instruments during the test invalidates the comparison

   ↓
2. 2

   Capture the order book ten times at equal intervals

   Five figures per snapshot plus the largest order

   ↓
3. 3

   Calculate the imbalance for each snapshot

   Five levels on each side, using the formula above

   ↓
4. 4

   Run a hypothetical market buy through the levels

   The volume-weighted average price and its difference from the best ask

   ↓
5. 5

   Compare the result with the next snapshot

   Whether the large order remained and where the average price moved

Diagram 4. The procedure for a practice test using ten snapshots. Source: a synthetic procedure based on the imbalance formula and the calculation of the average execution price.

Then consider one question. Among the snapshots with a strong imbalance, in how many cases did the average price actually move in the predicted direction by the next snapshot? You calculate your own figure, and it is more useful than anyone else's: it was obtained for your instrument, during your trading hours, and using your interval between snapshots. It also reveals how well conclusions drawn “from the picture” hold up in that specific market.

When Order Book Trading Can Be Misleading

The first trap is trusting the visible volume. A staff report from the Federal Reserve Bank of New York states this plainly:

“actual depth may effectively be lower than what is posted in the limit order book”

• Federal Reserve Bank of New York, Staff Report 827, https://www.newyorkfed.org/medialibrary/media/research/staff_reports/sr827.pdf

The distortion works in both directions. The same report notes the opposite as well: some participants do not display their full volume, so posted depth sometimes understates the true depth. The practical conclusion is the same: a figure from the order book is an estimate, not a measurement.

The second trap is the time horizon. The authors of the queue imbalance study tested whether the imbalance had meaningful predictive power and found a statistically significant relationship that was more pronounced for instruments with large price increments. However, they were referring only to the move up to the next tick:

“provides significant predictive power for the direction of the next mid-price movement”

• Gould, Bonart, “Queue Imbalance as a One-Tick-Ahead Price Predictor in a Limit Order Book,” arXiv, https://arxiv.org/abs/1512.03492

The third trap is fragmentation. A single instrument may trade on several venues, while your trading platform's order book shows only part of the queue. The fourth is speed: by the time a retail order reaches the exchange, the picture may already have changed.

Who This Approach Is Suitable For

Suitable if you

• Want to calculate the entry price before submitting an order rather than afterward
• Trade a volume large enough to be noticeable in your instrument
• Are prepared to keep a snapshot table and compare your conclusions with what actually happened
• Need a way to choose between a limit order and a market order

Not suitable if you

• Are looking for a way to guess the direction from the bars
• Hold positions for days or weeks
• Trade an instrument for which your platform's order book shows only part of the market
• Are willing to enter after a large order without checking whether it remains in the next snapshot

Risks and Limitations

Reading the order book does not provide an edge by itself and does not protect you from losses. These are the method's honest limitations.

• The tool does not guarantee a profit. Calculating the average price tells you the cost of entry, not the outcome of the trade.
• An imbalance signal is a reason to take a closer look, not an instruction to open a position.
• The market is volatile, and trading involves the risk of losing capital. During a rapid move, an order may be executed far from the expected price.
• Past results do not guarantee future results. The proportion of successful cases in the practice test does not carry over to future snapshots.
• The method is limited to a horizon of seconds and minutes and is unsuitable for position trading.
• Visible volume in the queue is not the same as available volume. Some orders are hidden, while others will be canceled before your trade.
• None of the calculations in this article replaces position sizing. The article on risk per trade separately explains how to calculate the acceptable loss on a trade.

Disclaimer. This is not individualized investment advice. The market is volatile, and trading involves the risk of losing capital. Past results do not guarantee future results.

Sources

• SEC, Trading Basics: order types and the absence of a guaranteed price for market orders, https://www.sec.gov/files/trading101basics.pdf
• FINRA, Order Types: why the execution price may differ from the price you saw, https://www.finra.org/investors/investing/investment-products/stocks/order-types
• FINRA, Rule 5310 Best Execution and Interpositioning: price, volatility, and relative liquidity as factors affecting best execution, https://www.finra.org/rules-guidance/rulebooks/finra-rules/5310
• FINRA, Stop Orders: Factors to Consider During Volatile Markets, https://www.finra.org/investors/insights/stop-orders-factors-consider-during-volatile-markets
• Moscow Exchange: execution of a market order against existing levels, https://www.moex.com/a2798
• Moscow Exchange: limits on market order aggressiveness and cancellation of the remainder, https://www.moex.com/a8519
• Moscow Exchange: WLIM orders and volume-weighted average price control, https://www.moex.com/n93802
• Binance help article: definition of slippage and a numerical calculation, https://www.binance.com/en/support/faq/what-is-slippage-01f6dd67d54e4dca902914700818e739
• Federal Reserve Bank of New York, Staff Report 827: actual depth may be lower than displayed depth, https://www.newyorkfed.org/medialibrary/media/research/staff_reports/sr827.pdf
• CFTC, JPMorgan Chase order dated 29.09.2020: the manipulator's reliance on traders who read the order book balance, https://www.cftc.gov/media/4826/enfjpmorganchaseorder092920/download
• Cont, Kukanov, Stoikov, arXiv: order flow imbalance and market depth, https://arxiv.org/abs/1011.6402
• Gould, Bonart, arXiv: queue imbalance as a predictor of the next mid-price movement, https://arxiv.org/abs/1512.03492
• TradingView, Strategy properties: slippage and commission parameters in the strategy tester, https://www.tradingview.com/support/solutions/43000628599-strategy-properties/

Frequently Asked Questions

How should a beginner use the order book?

Start with two figures: the spread and the total depth across five levels on each side. That is enough to determine whether entering with a market order is expensive right now. Add the imbalance and average price calculation once you are comfortable with the first two figures.

What is a limit buy order?

It is an instruction to buy at the specified price or lower. It joins the queue on the bid side and waits for a seller. The price is guaranteed, but execution is not: if the market moves higher, the order may remain unfilled.

Why was my order executed at a different price from the one I saw?

A market order takes volume from several successive levels. You buy the first part at the best ask and the remainder at worse prices. The average price is higher than the starting price, and this difference is called slippage.

Can I trade using the order book without the time and sales feed?

Technically yes, but in practice it is inconvenient. The order queue shows intentions, and without the time and sales feed, you cannot tell whether an order was filled or canceled. Checking the next snapshot partly closes this gap, but the time and sales feed provides greater accuracy.

How can the order book be used when trading on longer timeframes?

Almost not at all. An order book snapshot lasts for seconds, while a decision based on a daily chart is made on a different timescale. Its only use is to estimate the entry price once the decision has already been made.