Executing size: icebergs, TWAP and the OTC desk

Small orders rent the market’s prices; large orders change them. The moment your size becomes a meaningful fraction of the visible book, you stop being an observer of the price and become an input to it — and the market charges for that promotion. The charge is called market impact: the movement your own order causes, paid by you, at your expense, in both directions. Buying pushes the price up while you buy; selling pushes it down while you sell. It is slippage’s bigger sibling, and unlike fees it grows nonlinearly with size.

Everything institutions know about execution is a response to one problem: how to move size without announcing it. The tools — slicing, icebergs, time-weighted execution, OTC desks — are all camouflage of one kind or another.

Market impact, made concrete

Suppose you need to sell $200,000 of a coin whose book shows $30,000 of bids within 0.5% of the mid. A single market order eats those bids in milliseconds and keeps descending until it finds another $170,000 of willing buyers — perhaps 2% or 3% lower. On $200,000, a 2.5% average impact is $5,000, vaporized. Worse, the descent itself is a signal: other participants watch the book get swept, infer a motivated seller, pull their bids, and front-run the rest of your order downward. Size executed naively does not just pay the book — it teaches the market to charge more.

Slicing: the manual approach

The first remedy is the obvious one: break the order up. Sell $10,000, wait for the book to refill, sell again. Each slice is small enough to ride the spread rather than smash it. Slicing works, and for moderate size it is often all you need — but it has two costs. It takes time, during which the price can wander away from you (this is timing risk: impact and patience trade off against each other). And done by hand, with predictable size at predictable intervals, the pattern becomes legible — other traders can spot a clockwork seller as easily as a wall of asks.

Icebergs, conceptually

An iceberg order automates the camouflage on the book itself: the order has a large total size but displays only a small tip — show 500 units, hide 9,500. As the visible portion fills, the order automatically replenishes it from the hidden reserve. The book never reveals the true size, so the market is not warned. The limits are worth knowing: sophisticated participants detect icebergs by watching a level refill repeatedly after every fill, and the hidden portion typically waits behind visible orders at the same price. Obsidiate’s terminal does not offer iceberg orders — they belong to the conceptual toolkit here — but understanding them changes how you read books elsewhere, and explains why a level that keeps refilling deserves more respect than its displayed size suggests.

TWAP: trading the clock instead of the book

Time-weighted average price execution slices an order evenly across a chosen window: $120,000 to buy over two hours becomes $10,000 every ten minutes, mechanically. The goal is humble and useful — stop trying to pick the moment, accept the period’s average price, and keep each slice small enough to avoid impact. The trade-off is the same timing risk as manual slicing, plus the predictability of a metronome: anyone who spots the rhythm knows another slice is coming. Institutional desks add randomization and volume-awareness for exactly that reason. As a concept, TWAP teaches the central lesson of size: spreading an order through time converts impact cost into timing risk, and which cost is cheaper depends on how fast the market is moving against you.

When a block belongs on the OTC desk

Past a certain size, the order book is simply the wrong venue — every on-book technique still leaks information and still pays impact, just more slowly. Over-the-counter desks exist for this case: you request a quote for the full block, the desk prices it as a single all-in number, and the trade settles off the book. The market never sees the order, so it cannot move against it. The price will sit slightly outside the touch — the desk charges a spread for absorbing your risk — but for genuine size, one known, quoted cost beats an unknown and usually larger impact bill. Obsidiate’s OTC desk works this way: minimum ticket of $50,000, an all-in quoted spread with no separate fees to reconstruct, T+0 settlement, and segregated accounts. The arithmetic from our earlier example makes the case: $5,000 of estimated impact on a $200,000 sale is 2.5%, and an all-in OTC spread comfortably inside that number is not a convenience — it is the cheaper execution.

Key takeaways

• Market impact is the price movement your own order causes; it grows nonlinearly and is paid invisibly.
• Naive size also signals intent — a swept book teaches other traders to pull liquidity and front-run the remainder.
• Slicing and TWAP convert impact cost into timing risk; icebergs hide true size behind a small displayed tip.
• A price level that keeps refilling after fills is probably bigger than it looks — read books accordingly.
• Compare order size to visible depth before executing: small trades normally, moderate size slices, true blocks go off-book.
• OTC desks price the whole block as one all-in spread — on Obsidiate, from $50,000 per ticket with T+0 settlement.