Liquidity is used as though it named a single quantity that a market has more or less of. It names three, they are measured differently, they fail at different moments, and a market that scores well on the one you looked at can be the one that costs you the most.
The number quoted everywhere is volume, meaning how much traded over some period. It is a record of the past. Liquidity is a property of the present: what it would cost you to trade, right now, at your size. The two are correlated and they are not the same, and the cases where they diverge are precisely the cases that hurt.
A market can produce enormous volume through a small amount of capital changing hands repeatedly, which inflates the figure without adding any capacity to absorb a real order. It can also show modest volume while resting size sits patiently in the book, which is the opposite situation and looks worse on the metric everybody reads.
The practical version is that volume tells you a market was active and tells you nothing about whether you can get out of it. Traders who select instruments by volume have selected for attention, which is a reasonable thing to want and is not what they think they measured.
Volume is what happened. Liquidity is what you could do now. Only one of them is on your screen, and it is the other one that determines your fill.
Separating them is the whole of the subject, because each has a different cause and a different remedy.
The spread between the best bid and the best offer is the cost of buying and immediately selling, and it is the most visible of the three. It is also the least informative on its own, because it describes only the top of the book and says nothing about what happens beyond it.
How much rests within a given distance of the touch. This is the dimension that decides what a real order costs, it varies by orders of magnitude across the day, and it is invisible in every summary statistic commonly displayed.
After a large order consumes depth, the price recovers and the book rebuilds at some speed. In a resilient market this happens in seconds and the trade leaves little trace. In a fragile one the gap persists, which means the next order pays for the previous one, and a sequence of trades costs far more than the same size at once.
The spread is the number most platforms display and the one traders use to compare instruments. It is genuinely informative about very small orders and it can be actively misleading about everything else, because a market maker can quote a narrow spread on a tiny size and withdraw it the instant anything larger arrives.
The result is an instrument that appears excellent and behaves badly. The quote is tight, the top of book holds almost nothing, and a normal order walks several levels immediately. The measurement was accurate and it measured a quantity that does not scale to the size being traded.
The fix is to stop comparing spreads and start comparing what a fixed size costs. Take the size you actually trade, compute what it would pay to execute right now against the visible book, and compare that across instruments and across hours. It is the same effort and it measures the thing that determines your cost.
The book shows what participants have chosen to show. Hidden and iceberg orders mean some resting size is invisible, and conditional interest that has not been placed at all is invisible by definition: a participant willing to buy at a price they have not quoted is liquidity that exists and cannot be seen.
This cuts both ways and traders usually only notice one direction. On a well-traded instrument, the true depth is often better than displayed, because size appears when a price becomes interesting. On a thin one, the displayed depth can be worse than it appears, because the same order is posted on several venues by one participant who will cancel the others the moment one fills.
Nothing resolves this from the outside. What can be done is to stop treating the displayed book as an inventory and start treating it as a lower bound with unreliable error bars, which changes how much confidence a depth reading deserves.
There is one test that costs almost nothing and settles the question for the instruments you actually trade. Send a small order and watch what happens to the level it consumed: if it refills within seconds at the same price, there was more behind the quote than the quote showed, and the displayed depth is a floor. If the level stays empty and the spread widens, the displayed depth was the whole of it, and every size calculation you make on that instrument needs to assume the book is exactly what it says. The same test repeated at different hours tells you something no data provider will sell you, which is how the answer changes across the day on the specific pairs you hold.
Liquidity is supplied by participants who profit from providing it and who withdraw when providing it stops being profitable. That is not a moral failing, it is the arrangement, and it means supply is lowest precisely when uncertainty is highest.
Quoting into an announcement whose outcome you cannot price is offering a free option to anybody who knows more. Market makers widen or step back in the minutes before, so the book thins ahead of the event rather than during it.
A participant quoting both sides in a rapidly moving market is systematically filled on the wrong one. The rational response is to widen until the move settles, which removes depth at the moment orders most need it.
Nothing is wrong; there are simply fewer participants. The same nominal order is a much larger share of what is available, and an order sized for the busy session behaves like a much bigger order overnight.
The same asset trades in several places, and the depth is divided among them. Aggregate depth across all venues can be excellent while the venue you are on holds a fraction of it, and your order only ever meets the book in front of it.
This produces a specific and common error: reading an aggregate figure from a data provider, sizing a position against it, and then executing into a single venue that never had that depth. The aggregate was accurate and it described a market you were not trading in.
It also explains price differences that look like arbitrage and are not. A gap between venues that is smaller than the cost of moving assets between them is not an opportunity, it is the normal state, and the cost of that movement is the width the gap can sustain.
Depth that exists somewhere else is not depth you can trade against. Your order meets one book, and it is the only one that matters to your fill.
None of this requires special data. Three measurements, taken repeatedly on the instruments you trade, produce a picture that no summary statistic gives you.
First, the cost of your actual size, computed against the visible book and recorded at different hours. This single number replaces the spread as your comparison metric and it is the one that predicts your fills. Second, how long the book takes to rebuild after a large trade, which is the resiliency dimension and can be estimated by watching what happens after a visible sweep. Third, your own realised cost, comparing the mid price when you decided against your average fill, which captures everything the first two miss.
The third is the most valuable and the least done. It is the only measurement that includes what your own order did to the market, and traders who keep it for a month generally discover that their instrument selection and their trading hours matter more to their results than their entry criteria do.
A position is not risky only because the price might move. It is also risky because you might not be able to leave it, and those are separate exposures that get merged into a single stop-loss figure that only accounts for the first.
The usable rule is to size against depth as well as against account. If closing your position would consume a meaningful share of what is normally available, then your exit is going to move the price against you by an amount that has nothing to do with your analysis, and the stop you set is a number you will not receive.
This is the mechanism behind an experience most traders have had without naming it: a position that behaved reasonably while it was working, and cost far more than expected to close when it stopped. The market did not turn hostile. The size was always too large for it, and that only became visible on the way out.
No. Volume records how much traded in the past; liquidity describes what your order would cost now. A market can generate large volume from a small amount of capital turning over repeatedly, which inflates the figure without adding any capacity to absorb a real order.
Because the spread describes the top of the book and your order was larger than what sits there. A narrow quote on a tiny size is easy to offer and easy to withdraw. Compare instruments by what your actual size would cost to execute, not by the quoted spread.
How quickly the book rebuilds after a large order consumes depth. You see it by watching what happens after a visible sweep: in a resilient market the price and the depth recover in seconds, in a fragile one the gap persists and the next order pays for the previous one.
Treat it as a lower bound with unreliable error bars. Hidden and iceberg orders mean some resting size is invisible, and interest that has not been quoted is invisible by definition. On thin instruments the reverse also happens: the same order posted on several venues inflates apparent depth.
Because quoting into an event whose outcome cannot be priced offers a free option to anybody better informed. Market makers widen or withdraw in the minutes before. The move afterwards is what everybody watches; the disappearance beforehand is what actually costs money.
Only if you can reach it. Your order meets the book in front of it, so depth sitting on other venues is not depth you trade against. Sizing a position against an aggregate figure and then executing on one venue is a common and expensive mistake.
It should cap it. If closing would consume a meaningful share of normally available depth, your exit will move the price against you regardless of your analysis, and the stop you set is a number you will not receive. Size against depth as well as against account.
Your own realised cost: the mid price at the moment you decided, compared with your average fill, plus the fee. It captures the spread, your own impact and everything that moved while your order travelled. Keeping it for a month usually reorders which instruments and which hours are worth trading.