Iceberg Orders: The Liquidity You Cannot See

The offer that keeps coming back

Only 100 units appear for sale. Buyers trade 100, then another 100, then another. The offer keeps returning at the same price. After 800 units trade, the screen still shows 100. The initial displayed size did not describe all the liquidity that could become available.

An iceberg reveals a portion of a larger order. Native icebergs are managed by the exchange; synthetic ones use a participant’s software to submit fresh portions. Traders use them to reduce disclosure of their intended size. The presence of hidden size does not itself establish informed trading or manipulation.

Look for a sequence, not a screenshot

Zotikov and Antonov study CME order-level messages. Their native-iceberg detector uses executions exceeding an order’s displayed quantity and refreshes associated with the same order identifier. Their synthetic detector links new orders arriving shortly after a previous portion trades.

These are feed-specific methods. Repeated replenishment, similar portion sizes, and traded quantity far above an earlier display are clues, not universal proof. Independent traders can refill the same level; delayed or aggregated updates can resemble hidden liquidity. A synthetic trader can also vary price and portion size.

Research: CME iceberg detection — Zotikov and Antonov

Detection does not reveal the remainder

Even if 800 units have executed, you do not know whether 100 or 10,000 remain. The CME paper estimates total size using a survival-analysis method that includes partially executed orders later canceled. Ignoring these incomplete orders would distort what the model learns.

The useful output is an uncertain estimate, updated as evidence arrives. Test false positives and performance on later data. A detector calibrated on CME messages does not automatically work on Binance or Polymarket, whose available feeds and order semantics must be checked separately.

Hidden liquidity can help you execute

Frey and Sandås find that hidden liquidity is associated with changes in order flow and price impact in their equity-market sample. A replenishing seller can absorb buying that otherwise would have moved the price through several levels. That may help a buyer seeking size.

Other research finds that both informed and liquidity-motivated traders use icebergs, and that repeated replenishment helps the market discover them. Absorption describes trading without much price movement; it does not prove one hidden parent order or predict which way the next move will go.

Research: hidden liquidity — Frey and SandåsResearch: who can see the iceberg’s peak?

Change the execution question

If you want to buy, a persistent sell offer may provide liquidity within your price limit. If you are buying a breakout, the same offer weakens the assumption that visible selling is almost exhausted. Keep size and maximum price explicit instead of treating each refill as permission to trade more.

Watch what happens when replenishment stops: the seller may be finished, may have canceled, or the feed may have paused. Require usable data and fresh evidence. Hidden liquidity can disappear, so neither a bounce nor a breakout is guaranteed.

Write down what would disprove your detector

Log timestamps, trades at the level, displayed depth, refill delays, and subsequent price changes. Compare decisions with and without the detector, including fees and missed fills. Keep the detection rule fixed during evaluation on new data.

Treat iceberg detection as a research input to execution. The first practical skill is learning where the visible book stops being a complete description.