Displayed Depth and What Actually Fills

An order book displays resting orders. That is an upper bound on available liquidity rather than a measurement, and the difference is larger than most participants assume. For a reference point, a crypto OTC trading platform documents its settlement mechanics rather than agreeing them per trade.

Why it overstates

Cancellation. Much resting depth is placed by market makers who withdraw when conditions change, which is precisely when you want to trade.

Duplication across venues. The same capital appears on several books simultaneously. Summing depth across venues double counts.

Layering. Orders placed at multiple levels by the same participant, only one of which they intend to fill.

The measurement

Execute at a known size and compare the average fill price against the mid at submission. The difference is the effective cost.

Repeat at several sizes and you have a depth curve for that venue and asset, which replaces an assumption with a number.

In calm conditions, displayed depth typically overstates what fills by a factor of two to three at meaningful size. During volatility, considerably more. Funds face the same question with an extra reporting layer, which is what a provider serving funds and family offices is structured around.

The threshold for a desk

Lower than most assume, because of the above.

The practical test: sum resting liquidity within half a percent of mid. If your order exceeds roughly a tenth of it, request a desk quote before deciding. Above half, use a desk.

The concentration effect

Market making consolidated after several cycles. A small number of firms now provide much of the depth on major venues.

Spreads narrowed and became more stable, which is a genuine improvement.

Behaviour during volatility became more correlated. Similar models widen at similar moments, so depth disappears more uniformly than it used to.

What this means for execution

Avoid market orders during volatile periods. The realised cost bears little relationship to what the book showed a second earlier.

For anything time sensitive at size, price through a desk that will quote firm.

The data point worth collecting

Effective spread against mid, per execution, by venue and size.

Ten executions give a real picture. It is the only information that reflects your actual flow, and no published figure substitutes for it. Whatever you conclude here, the balance you actually trade belongs at a crypto exchange with published fees rather than wherever the interface was friendliest.