What Is a Dark Pool?
Michael Whitman
·
Trading since 2017, full-time since 2020
·
YouTube creator, 137K subscribers
·
Published 2026-09-15
·
Updated
Dark pool is a private trading venue where orders are not displayed to the public before execution, so a large buyer or seller can transact without revealing intent in advance. The completed trade is still reported afterwards, meaning the concealment covers timing rather than the fact of the trade.
A dark pool is a venue where the order book is not published. Everything unusual about it follows from that one design choice, and almost everything said about it online does not.
How it works
Orders rest without being displayed. On a public exchange a resting order is visible; in a dark pool it is known only to the venue until it matches.
Execution is still reported. Once a trade happens it is published to the consolidated tape like any other, so the volume appears in the public record.
Pricing usually references the public market. Many pools match at the midpoint of the public bid and ask, so the lit market sets the price and the dark venue supplies the meeting place.
The problem they exist to solve
A very large order cannot be shown. Displaying an intention to sell a million shares tells everybody what is coming, and price moves before the order is filled.
That effect is called market impact. It is a real cost borne by the institution and, since most institutions manage other people’s money, ultimately by the savers behind it.
A dark venue reduces it. The order finds a counterparty without first announcing itself, which is the entire mechanism and the entire justification.
A worked example
Suppose a fund needs to sell a position worth many days of average volume. Placed openly, the first slice moves price against the remaining slices.
Even an ordinary-sized trade in a liquid market pays the bid-ask spread on every round trip, and that is the cheapest case there is.
A large order pays far more than that. It consumes the displayed size at each level and then the next, so the average fill is worse than the price that was on screen when it started.
The dark pool is an attempt to cap that. Matching at the public midpoint avoids walking the book, and whether it succeeds depends entirely on whether a counterparty of matching size is there.
What the prints do and do not tell you
A dark-pool print is a completed trade. It records size, price and time, after the fact.
It does not record direction. The tape does not say which side initiated, so “dark pool buying” is an inference from the print, not a field in it.
Nor does it record who. Counterparty identity is not published, so any claim about which institution traded is reconstruction rather than data.
And the size of a print is not the size of an intention. A large order is usually split, so one print may be a fragment of something much bigger or the whole of something small - and nothing in the record distinguishes the two.
The original data
The S&P 500 closed up on 53.1% of trading days between 1950 and 2026 — on any single day, not far from a coin flip.
Set that near coin flip against the trading signals built on dark-pool prints. A print reports something that already happened, and in a market whose daily direction is that close to even, information about the last trade has a short shelf life.
And acting on each print is expensive. Every round trip pays the spread and any commission, so a feed producing dozens of prints a day produces dozens of chances to pay that on information the market has already seen.
Not the same as internalization
Retail orders filled by a wholesaler are also off-exchange, and that arrangement gets called a dark pool in a lot of commentary. It is a different thing.
A wholesaler fills your order against its own inventory. There is no pool of resting institutional orders involved, and the fill typically references or improves on the public quote.
A dark pool matches two participants with each other. Its purpose is bringing large opposing orders together without displaying either.
The confusion matters because the criticisms differ. One raises questions about order routing incentives; the other raises questions about price discovery moving off the lit market - and treating them as one topic makes both arguments worse.
Why they are regulated rather than secret
They are registered venues. Operators file with regulators, report trades to the public tape, and are subject to the same trade-reporting rules as lit exchanges.
Their volumes are published. Aggregate activity by venue is disclosed regularly, so the share of trading that happens off-exchange is a public figure rather than a rumor.
Enforcement actions exist and are public. Where operators have misrepresented how their pools worked, the cases are on the record and readable.
Which is worth holding onto. “Not displayed before execution” and “unregulated and hidden” are very different claims, and only the first one is true.
How this connects
The public alternative is the stock exchange. There, resting orders sit visibly at the bid and the ask, and every participant can see the queue a large order would have to work through.
Lit venues offer their own partial concealment. An iceberg order shows only a slice of its size at a time, which tackles the same market-impact problem without leaving the exchange.
Institutions often combine both with algorithmic trading, slicing a parent order across lit and dark venues and measuring the result against VWAP to judge whether execution was any good.
Slippage is the cost all of this is fighting. Every price level consumed on the displayed book pushes the average fill further from where the order began.
The retail arrangement described above runs through a market maker, and the payment for order flow behind it is a separate debate from anything happening in a dark pool.
High frequency trading firms are part of the story too, because concern that fast traders could detect resting institutional orders is one reason venues restrict who they admit.
And prints do not settle differently. A dark trade moves through the same settlement cycle as any other once it has matched.
When it fails
The characteristic failure is treating prints as a signal. A feed of large off-exchange trades looks like privileged information, and it is the opposite - it is a record of decisions already executed.
The interpretation then supplies what the data lacks. Direction, motive and identity are all absent from the print, so they get filled in by whoever is reading it, and a confident story is built on three guesses and one fact.
A second failure is assuming large means informed. Index rebalancing and redemptions produce enormous trades with no view attached.
A third is expecting retail access. These venues serve institutional flow; a retail order does not reach them as a resting order.
A fourth is reading off-exchange share as a health measure, when it mostly reflects how institutions choose to execute.
The question: The reading is clear. What does it leave out?
And a fifth is believing price is set there. Most pools reference the public quote, so the lit market is doing the pricing and the dark venue is doing the matching.
Related
Liquidity covers why a large order moves price at all. Order book covers the displayed alternative. And volume covers what the prints actually add to.
This is the most mythologized piece of market plumbing there is. The actual purpose is boring: a pension fund selling a very large position does not want everybody to know before it is done. What frustrates people is that the prints arrive after the fact, and a record of something that already happened is a poor basis for a prediction.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.