Market Depth and Order Books: Read Size Beyond the Best Price

What the columns mean

An order book groups buying interest on the bid side and selling interest on the ask side. Each row associates a price with a quantity. The best bid and ask form the top of the book; additional rows show deeper levels. Spotware documents depth views and volume-weighted average prices, but the available view depends on the platform and the underlying data source.

Read the quantity unit before using a row. Four contracts, four lots and four units are not interchangeable. Also identify whether the display is an exchange book, aggregated liquidity available through a broker, or a simulated feed. A broker's visible depth is not a consolidated inventory of all buying and selling interest worldwide.

A four-contract order through two levels

Imagine an instrument with the following ask prices. This deliberately simple example assumes that displayed quantities remain available and there are no extra execution restrictions:

Hypothetical ask-side depth

Ask price	Available contracts	Used by a market buy of four
100.10	2	2
100.20	3	2
100.30	5	0

The average entry price is (2 × 100.10 + 2 × 100.20) ÷ 4 = 100.15. The best ask was 100.10, but it covered only half the order. The 0.05 average difference is a price distance. To express it in money, multiply by four contracts and the instrument's currency value per price unit.

A buy limit of 100.10 would restrict the acceptable execution price. Under these assumptions it could immediately obtain at most two contracts at that level. What happens to the other two depends on the order's lifetime and filling policy. See time in force and partial quantities before assuming the remainder stays active.

What depth leaves unknown

The screen and the order do not arrive at the matching system simultaneously. Orders ahead of yours may consume visible liquidity; resting interest may be canceled; additional interest may appear. Some markets also permit forms of undisplayed quantity. A screenshot therefore cannot prove that a later order should have filled at its displayed price.

A limit order resting at a traded price is not automatically first in the queue. The market's matching rules, earlier orders and the actual aggressing quantity all matter. This is why a candle touching a limit price is weaker evidence than the relevant order and transaction records.

Who fills first at one price? A scoped FIFO example

For this separate exercise, EDU-FUT is a fictional futures instrument with one bid price of 100.00, three accepted outright buy-limit orders, whole-contract quantities and a declared CME Globex-style FIFO (F) matching algorithm. Assume no implied orders, top-order or lead-market-maker treatment, hidden quantity, cancellations or intervening arrivals. A sell order for five contracts then arrives at that bid. This is a teaching sequence, not CME market data, an observed broker route or a TradeCopier result.

CME's algorithm definitions say that FIFO fills same-price resting orders in time order and that increasing an order's quantity re-queues it. Its matching-steps documentation also lists pro-rata and mixed algorithms. Confirm the current instrument's algorithm from its futures security definition; do not apply this FIFO exercise to every futures contract, spread, venue or broker interface.

In the unchanged branch, the five-contract sell fills A's three first and two of B's four; C receives none. Nine contracts were resting before the match and four remain: B two plus C two. In the alternative branch, B's quantity increase loses its earlier time priority, so the queue is A, C, B. The same sell fills A three and C two; B receives none. Ten contracts were resting before that independent match and five remain with B. These branches are not consecutive trades, and a different amendment or matching algorithm can produce a different allocation.

Download the six-row synthetic queue ledger (CSV) to check each priority time, allocation and remainder. The fill quantities sum to five in each branch. An aggregate market-by-price level cannot reveal whether the five contracts went to A and B or A and C. CME distinguishes aggregated market-by-price data from market-by-order data, in which anonymous OrderID and PriorityID can help reconstruct individual priority. Even an order-level feed does not guarantee a later order will fill: the book and your priority can change before matching, and actual order acknowledgments and execution records decide the outcome.

For a copied-order investigation, record the exact destination venue or broker, instrument, algorithm when applicable, accepted order ID, amendment history, price, quantity, timestamps, partial fills and final position for each account. Do not infer a follower's queue position from the source account's depth display. This example explains one exchange matching rule; it does not establish that TradeCopier supports a particular futures route or that an eligible instruction received a fill.

Reconcile the result rather than the screenshot

Keep the requested quantity, order acknowledgment, every partial deal, any cancellation and the final position. Add filled quantities before calculating their weighted average. For example, two deals of two contracts each are one four-contract result; averaging their prices without quantity weighting works here only because their sizes are equal.

Across copied accounts, compare each account's own fills and available instrument. Different books, arrival times or requested sizes can create different results even from the same source instruction. The existing partial-fill guide extends that reconciliation task. None of these observations establishes a universal fill guarantee or platform compatibility.

Calculate a scoped, visible order-book imbalance

An imbalance compares displayed bid and ask quantities in a specified view. For this exercise, use the first three price levels on each side of one invented venue, at one timestamp, with quantities in whole contracts. Define I3 = (B3 - A3) / (B3 + A3), where B3 is the sum of bid contracts at levels 1 through 3 and A3 is the matching ask sum. A positive value means more visible bid quantity in those levels of that snapshot. It does not identify the next price move, all available liquidity, participant intent, or an executable fill.

CME distinguishes market-by-price data, which aggregates quantities at each price level, from market-by-order data, which adds anonymous individual-order detail. Do not treat a price-level summary as a queue-position record. CME's liquidity-tool methodology also describes timestamped book updates after orders, trades and cancellations. Neither source defines this three-level I3 as a CME signal; the ratio and its three-level boundary are our declared teaching convention.

The downloadable example uses EDU-XYZ, an invented instrument with a 0.10 price tick. It is a synthetic single-venue price-level feed, not TradeCopier data, CME market data or a cTrader account. All nine rows record the same unit and feed scope. The event labels are part of our fictional script; two real snapshots alone would not reveal whether a size change came from cancellation, execution, replacement or another cause.

Three invented snapshots, each using exactly three displayed levels per side

Timestamp (UTC)	Bid levels, contracts	Ask levels, contracts	B3 / A3	I3	Best spread
14:30:00.000	8 + 6 + 4	4 + 5 + 3	18 / 12	6 / 30 = +0.20	100.00 - 99.90 = 0.10 = 1 tick
14:30:00.250	8 + 6 + 4	2 + 5 + 3	18 / 10	8 / 28 = +0.285714	100.00 - 99.90 = 0.10 = 1 tick
14:30:00.500	2 + 2 + 1	2 + 1 + 1	5 / 4	1 / 9 = +0.111111	100.20 - 99.80 = 0.40 = 4 ticks

At 14:30:00.250 our scripted event cancels two contracts at the best ask and records no trade. I3 rises from 20% to about 28.5714% while the top prices and spread stay unchanged. At 14:30:00.500 a new book has just nine visible contracts across both sides and a four-tick spread; the intervening event sequence is deliberately unspecified. Its positive 11.1111% imbalance is not evidence of ample executable size. Both CME's fill-quality analysis and the changing example show why one displayed depth number cannot stand in for a full execution assessment.

Download the nine-row synthetic snapshot CSV. Recalculate bid and ask totals from the three rows for each timestamp before using the ratio. If either side is missing, the feed timestamp is unavailable, quantity units differ, or the planned three levels are incomplete, label this exact calculation unavailable instead of silently treating a missing value as zero. If B3 + A3 is zero, the ratio is undefined.

Compare feeds and timestamps before interpreting the number

The scope of a depth view matters. Spotware documents Standard, Price and volume-weighted-average-price depth views in cTrader; these displays answer related but different questions. A cTrader view cannot be assumed to contain the same venues, orders, levels, units or timestamps as a CME price-level feed. Record the exact instrument, feed/provider, selected side and levels, quantity unit, time zone, snapshot timestamp and price increment beside every computed I3. A comparison across accounts is meaningful only after those fields are made comparable.

Displayed interest can be canceled or replaced before an instruction arrives. A market-by-price summary cannot establish an individual order's queue place; even more detailed order-level data does not reserve displayed size for a later incoming order. A quote that was fresh at capture can already be stale for a decision. Keep the event sequence and actual fills separate from the pre-trade snapshot.

Use a decision record when visible liquidity looks thin

The table below is a blank operating-policy example, not a buy or sell recommendation. Define any thresholds, permitted order types and who may approve an exception before seeing the outcome. The same snapshot can support different choices under different mandates; none of the rows claims that waiting, using a limit or sending no order will improve returns.

What to observe before choosing an action under a separately approved policy

Observable state	Evidence still needed	Possible rule to document
Displayed spread widens and size on the intended execution side is below the proposed quantity	Fresh timestamp, deeper levels, order size and permissible price boundary	A policy may wait for a fresh book, set a limit that can remain unfilled, or send no order.
One side or its timestamp is missing	Feed health and venue or broker trading status	Mark the cost estimate unavailable; a no-order rule can apply until the data is restored.
Two accounts show different depth or spread	Instrument identity, unit conversion, provider scope and capture times for both feeds	Review each destination independently rather than carrying one account's estimated fill to the other.
A narrow top spread appears with little visible quantity	Depth beyond the best price, arrival time and actual executed deals	A limit may bound price but has nonfill or partial-fill risk; a wait or no-order rule remains possible.

A later review should compare the chosen reference price and time with each actual fill, quantity, fee and final position. The original four-contract example above shows how a weighted fill can differ from the best displayed ask. This additional I3 worksheet measures a different thing: the balance of visible resting size in three selected levels. Neither metric alone explains a broker result or validates a TradeCopier feature. Preserve the original order and deal records before attributing a difference to thin liquidity.