Liquidity · Reviewed Aug 19, 2026 · 8 min read
Order Book Depth: The Metric That Actually Predicts Your Fill
The 0.1% and 2% depth bands tell you more about an exchange than any volume ranking ever will.
Reviewed by CryptoExID Editorial · Aug 19, 2026 · Editorial policy · how we make money
What depth measures
Order book depth is the dollar value of resting limit orders within a set distance of the mid price. It answers the only question that matters before you press buy: how much can this market absorb before the price moves against me? Volume answers a different and weaker question, namely how much traded yesterday.
Depth is quoted per side. Bid depth absorbs your sell; ask depth absorbs your buy. A healthy market keeps both sides roughly balanced. A book with $5 million of bids and $300,000 of asks is telling you something, and it is not something good.
The 0.1% band: cost for normal orders
Depth within 0.1% of mid, often written as plus or minus 0.1%, is the tightest standard band. It captures what a market order of ordinary size will actually touch. If BTC sits at $100,000, this band covers orders resting between $99,900 and $100,100.
On Binance, our 9.9-rated leader, the 0.1% band on BTC/USDT typically holds millions of dollars per side. A $50,000 market order disappears into that without a ripple. On a 6.8-rated book the same band might hold under $100,000, and the same order walks straight through it.
The 2% band: stress capacity
Depth within 2% of mid measures what happens under stress: a liquidation cascade, a headline, a whale in a hurry. It is the market's crumple zone. Two venues can look identical in the 0.1% band while one holds ten times the capital in the 2% band, and that difference decides who survives a flash move intact.
We treat the 2% band as a venue-quality signal rather than a cost estimate. If you are routinely trading 2% of the way into a book, you should be splitting orders or using an OTC desk, not reading comparison sites.
Why depth beats volume as a quality signal
Faking volume is nearly free: the exchange matches its own orders and the ticker climbs. Faking depth requires posting real orders that any stranger can hit. Anyone can lift those quotes at any moment, so fake depth carries real financial risk for the faker. Depth is therefore the harder signal to counterfeit and the one we anchor our subscore on.
The 2019 Bitwise analysis made this concrete: roughly 95% of reported BTC volume was fake, yet the honest venues were easy to identify by their books. Depth separated the real markets from the theater.
Reading depth like our raters do
Three checks cover most of it. First, absolute size: how many dollars sit within 0.1% and 2% of mid on the pair you trade. Second, balance: bids and asks should be the same order of magnitude. Third, persistence: depth that vanishes during volatile hours is decoration, not liquidity, which is why spread persistence is a separate input in our model.
In our current set, Binance 9.9, Bybit 9.6 and OKX 9.5 pass all three on the majors. Upbit at 9.1 is remarkable on KRW pairs specifically. Gemini at 7.6 is honest but noticeably thinner.
The limits of our measurement
We do not stream live books to CryptoExID. Our liquidity subscore is built from depth observed on the dataset date, plus spread persistence and venue tier. That makes it a fair ranking tool and a poor execution tool. For a specific large trade, open the venue's own depth chart and look at the actual band you will consume.
FAQ
What is a good 0.1% depth for BTC?
Top venues hold several million dollars per side within 0.1% of mid on BTC/USDT. If a venue shows less than $100,000 in that band on a major pair, treat it as thin and size your orders accordingly.
Why do you look at 2% depth if nobody trades that deep?
It measures stress capacity. During liquidation cascades and news spikes, the 2% band is what stands between a sharp move and a disorderly one. It is a venue-quality signal more than a cost estimate.
Can depth be faked like volume?
It is much harder. Fake depth means posting real orders that anyone can execute against, so the faker carries genuine risk. Spoofers do flash orders and pull them, which is why we also check spread persistence over time.
Is one-sided depth a red flag?
Usually. Heavily imbalanced books suggest either strong directional pressure or a market maker who has stopped quoting one side. Either way, expect worse fills in the thin direction.
Where can I see live depth data?
On the exchange itself, via its depth chart or API. CryptoExID publishes a scored subscore checked on our dataset date, not a live feed, and we recommend checking the actual book before any large order.