The mechanism

#The root idea

A price chart shows you one number per moment: the last trade. It throws away who initiated that trade, how much size was resting on either side of it, and what happened to all the orders that didn't trade. The order book and the tape are the record of that discarded information — and the discarded information turns out to be predictive.

Every trade in a modern electronic market happens because one side was passive (a limit order resting in the book, waiting) and the other was aggressive (a market order, or a marketable limit order, that crossed the spread to take that resting liquidity immediately). The passive side is providing liquidity — offering to trade at a stated price and waiting. The aggressive side is taking liquidity — paying to trade right now rather than wait. That asymmetry is the whole mechanism: a taker is revealing something by choosing to pay for immediacy instead of waiting for a better price, and the order book is a live record of how much of that revealed urgency is stacking up on each side. [Established] — this is the standard description of a continuous double auction, the mechanism underlying essentially every modern electronic exchange; see Market microstructure fundamentals for how an exchange actually implements it.

#The peer-reviewed finding this course is built on

The specific, tested claim is this: order flow imbalance — the net signed volume of aggressive buying versus aggressive selling, including order cancellations, over a short window — has a linear, measurable relationship with the price change over that same window, and the slope of that relationship is inversely proportional to the depth resting in the book. Thinner books move more for the same amount of order flow; deeper books absorb more before price gives way.

This isn't a trading-forum claim. It's Rama Cont, Arseniy Kukanov, and Sasha Stoikov, "The Price Impact of Order Book Events," Journal of Financial Econometrics 12(1), 2014 — peer-reviewed, tested against NYSE trade-and-quote data across fifty US stocks, robust across time scales and across the sample. [Established] — see Sources and provenance for the full citation. Read what it actually claims narrowly: it's a statement about short-horizon price impact from order-book events in the sample it tested. It is not a claim that order-flow reading generalizes to every instrument, every horizon, or every market regime, and this course doesn't stretch it into one.

#Why the information exists at all: adverse selection

The deeper "why" is adverse selection. A market maker posting a resting limit order doesn't know whether the next person to trade against it is a random liquidity need (a fund rebalancing, someone closing a position) or someone who has a genuine, current information edge about where price is about to go. If it's the latter, the market maker's resting order gets picked off — filled right before price moves against the price they just quoted. Market makers price this risk into the spread they quote and manage it by watching the shape of incoming flow: persistent, one-directional, size-escalating aggression looks more like informed flow than random two-sided noise, and market makers widen spreads or pull size when they see it. [Established] as a description of market-maker behavior generally — this is the standard adverse-selection account of why spreads exist and move, going back to the foundational Glosten-Milgrom sequential-trade model of market microstructure.

That's the mechanism this whole course teaches you to read, not exploit at the market maker's own speed: if informed order flow has a detectable signature in the shape of the tape and the book, and if that signature has a short-horizon relationship with price that's been measured and published, then reading the book and tape is a legitimate, evidenced information source — distinct from, and prior to, any specific trading strategy built on top of it.

#What this mechanism does not claim

It does not claim you can see who is trading or why — order flow imbalance is a statistical pattern in anonymized, aggregated data, not a window into institutional intent. It does not claim the edge is large, stable, or accessible to a retail participant at retail latency — Why this is hard to sustain solo covers exactly what changes when you can't see the flow at institutional speed or scale. And it does not license the retail "smart money concepts" framing that treats every order-block pattern as direct evidence of a specific institution's hand — that framing is addressed and rejected on its own terms in Order flow imbalance and absorption.