Blog
Blog
Every retail trader knows the word "spread." Almost none of them know which half of it they are paying, when it collapses to zero, and why that is the moment to be most careful.
Market microstructure is the study of the mechanics underneath price. Not what prices do over days and weeks, but what happens in the seconds before a trade executes: who is providing liquidity, who is taking it, how much of each, and what that distribution implies about what happens next. The vocabulary of this field is not jargon for its own sake. Each term names a variable that determines execution quality in ways that price charts do not capture.
Use this glossary as the companion map for What Is Market Microstructure?, Order Flow Imbalance Explained, and What Market Depth Measures. The definitions matter only because they change how a trader interprets live context.
A market at any moment has two prices, not one.
The bid is the highest price a buyer is willing to pay right now. The ask is the lowest price a seller is willing to accept. The spread is the difference between them. On BTC during liquid hours on a major exchange, that spread might be a few cents on a $60,000 asset. On a low-volume altcoin at 3am UTC, it can be 0.5% or wider.
The spread is not a fee. It is the cost of immediacy.
A trader who places a market order pays the spread: they buy at the ask or sell at the bid, always the unfavourable price. A trader who rests a limit order inside the spread collects it: they post a bid above the best bid or an offer below the best ask, and wait for the market to come to them. The first type of order is called a market order or taker order. The second is a maker order. The exchange charges takers and rebates or discounts makers because makers are providing a service: they are creating the inventory that takers need.
Crossing the spread intentionally signals urgency. Urgency has a price.
The spread is a snapshot at the top. Depth is the full picture.
An order book is a sorted list of all resting limit orders: bids at every price below the current best bid, offers at every price above the current best ask. The volume at each price level represents buying or selling interest that is not urgent enough to cross the spread but real enough to be posted. Depth is the aggregate of this resting inventory.
A deep book means large orders can execute with minimal price impact. Shallow depth means even a moderate order moves price significantly. During the FTX collapse in November 2022, books across crypto markets became so thin that a single sell order in the hundreds of thousands of dollars moved BTC by hundreds of dollars within seconds. That was not a price discovery event in any normal sense. It was a depth event: the visible inventory required to absorb selling had vanished.
Depth is also the first place manipulation appears. Orders posted and cancelled in milliseconds, walls of bids that disappear the moment price approaches them, identical order sizes refreshed repeatedly at the same level, all represent attempts to shape what the book looks like without genuine intent to trade. The book as displayed is not always the book as intended.
Transaction costs have three components. Most traders only count one.
Commission or fee is visible: the exchange charges a percentage. The spread is semi-visible: you can see the bid-ask and estimate the crossing cost. Market impact is invisible until after execution.
Market impact is the price change caused by your own order. A $10,000 buy order on a liquid instrument moves price by a negligible amount. A $5 million buy order moves price against you as it executes: early fills are cheap, later fills are more expensive because your order is consuming the available offers and the book is repricing as you go. By the time the last portion fills, the average execution price is worse than the price at order entry.
This is not slippage in the colloquial sense of prices moving for other reasons. This is the mechanical consequence of being large relative to the available depth. Institutional desks model this explicitly and split orders algorithmically to reduce the footprint of each individual execution. At sufficient scale, ignoring impact means paying a cost that exceeds any alpha a strategy generates.
Volume tells you how much traded. Order flow tells you who was in a hurry.
Every trade requires a buyer and a seller. What differentiates them is not the direction but the role: one crossed the spread, one rested. The one who crossed was aggressive. The one who rested was passive. Signed order flow tracks aggression directionally: buy-initiated trades (buyer crossed) versus sell-initiated trades (seller crossed). When buyers are significantly more aggressive than sellers over a rolling window, demand is outrunning supply at the current price, and price tends to follow.
Cont, Kukanov, and Stoikov formalised this in a 2014 paper in the Journal of Financial Econometrics (12(1):47-88). The result was not that signed order flow predicts price with certainty. The finding was cleaner: at horizons of seconds to minutes, the imbalance between aggressive buying and aggressive selling has statistically significant explanatory power for mid-price changes, beyond what current price itself provides. The chart is the lagging indicator. Order flow is the leading one.
Order Flow Imbalance (OFI) operationalises this: it quantifies the asymmetry between aggressive buy and sell volume at the top of the book. Sustained positive OFI means buyers are urgently consuming offers. The price has not moved yet. The force behind the move is already measurable.
Price is not a single number even when the book is quoted.
The mid-price is the arithmetic average of the best bid and best ask. On a book with a bid at $65,000 and an ask at $65,002, the mid is $65,001. Simple. The mid-price is used as the reference for calculating spreads, measuring price impact, and computing PnL in many quantitative frameworks.
The microprice is more sophisticated. It weights the mid by the relative volume at each side of the book: if there are three times as much resting bid volume as offer volume at the top, the microprice shifts toward the bid side, reflecting the imbalance in available liquidity. A symmetric book has a microprice equal to the mid. An imbalanced book does not.
In practice, a microprice significantly above the quoted mid suggests buy-side depth is thin and sellers are close: the next aggressive buyer will move price faster than the mid implies. The microprice is a real-time estimate of where the next trade is likely to occur. Strategies operating at sub-second frequencies use it rather than the mid because the mid ignores the very information it is constructed from.
The word "latency" suggests a technical concern. It is a business concern dressed in technical language.
Latency is the time between an event in the market and a system's response to that event. The event might be a large order appearing in the book, a price move on a correlated instrument, or a trade printing on a different venue. The response might be posting a new quote, cancelling an existing one, or sending an aggressive order.
Slower participants trade on stale information. A market maker quoting at a price that was accurate 50 milliseconds ago is offering a free option to anyone faster: they can buy the stale offer if price rose, or sell the stale bid if price fell. The price of being slow is adverse selection. Faster participants consistently pick off slower ones, not through skill, but through information advantage created entirely by time-to-response.
At the fastest tier of equity and futures markets this gap is measured in nanoseconds, enforced by co-location, custom network hardware, and FPGA execution. Crypto markets are slower, which is why they remain accessible to participants operating at millisecond rather than nanosecond scale. The dynamic is identical. The gap is just larger.
Resting a limit order sounds safe. It is not costless.
When a taker hits a resting order, one of two things is usually true: either the taker has no special information and is simply buying or selling for portfolio reasons, or the taker knows something about where price is going. The second case is adverse selection. You provided liquidity at a price that, moments later, was wrong. The taker chose to trade against you precisely because they knew something the book did not yet reflect.
Adverse selection is why market making is not a free lunch despite collecting the spread. A market maker earns the spread on uninformed flow and loses to informed flow. The ratio between them determines profitability. On instruments where the ratio of informed to uninformed order flow is high, market making is structurally unprofitable unless you can identify and route around the informed side.
This is also why orderbook analysis is not just about size and position. The source of order flow matters. Repeated aggressive orders from a single direction, concentrated at specific times, following correlated movements on other venues, suggest an informed actor rather than random noise. Distinguishing the two is the core problem of market microstructure analysis.
These terms do not belong to an academic specialty that has no application to actual trading.
Spread, depth, impact, mid-price, microprice, signed flow, adverse selection, latency: each names a quantity that is continuously changing, that affects every execution, and that price charts completely discard. A trader who does not know what adverse selection is cannot recognise when they are experiencing it. A trader who does not distinguish signed order flow from volume is counting something that does not directionally distinguish buyers from sellers.
The firms that operate systematically in these markets have not simply built better models. They have built representations of what is happening that contain more relevant information than the representations available to everyone else. Vocabulary is where that divergence starts. The mathematics follows from it. The engineering follows from that.
The spread you see is the summary. What this vocabulary opens up is everything the summary discards.
DepthSignal computes signed order flow, depth metrics, and imbalance signals across many live exchanges, normalised and continuously updated.
No. The vocabulary is useful anywhere execution quality, liquidity, pressure, or market impact matters.
Spread is the simplest starting point because it turns the abstract idea of liquidity into an immediate execution cost.
Price shows the result. Order flow shows which side applied pressure and whether that pressure had enough liquidity behind it to matter.
Cont, R., Kukanov, A., and Stoikov, S. (2014). The price impact of order book events. *Journal of Financial Econometrics*, 12(1):47-88.
Hasbrouck, J. (2007). *Empirical Market Microstructure*. Oxford University Press.
Harris, L. (2003). *Trading and Exchanges: Market Microstructure for Practitioners*. Oxford University Press.