Bid–ask spread

Ela opens the market for BTC against dollars. The order book, a list of waiting buy and sell orders, shows a best bid of 59,990 USD and a best ask of 60,010 USD. To buy immediately, she accepts the seller’s higher price. To sell immediately, she accepts the buyer’s lower price. The gap is 20 USD per BTC. Bid and ask describe the sides of the waiting offers, not the buttons Ela is pressing. This is our simplified example; a real trade also depends on how much bitcoin remains available at each price. [1]

The midpoint between the offers is 60,000 USD. It is the average of the best bid and ask, useful for comparison but not a promise that Ela can trade there. Relative to this midpoint, a spread of 20 USD is about 3.33 basis points. A basis point is one hundredth of a percentage point, so that is roughly 0.0333%, not 3.33%. For two relative figures to mean the same thing, they need the same denominator: the amount used as the base in the division. Dividing by the bid differs slightly from dividing by the midpoint. Understanding the dollar prices comes first for an ordinary purchase; these units help with more precise market comparisons. [1]

In our example, Ela buys 1 BTC for 60,010 USD, which is 10 USD above the 60,000 USD midpoint. If she immediately sells at the unchanged bid of 59,990 USD, she receives less than she paid. The difference is 20 USD before fees, not 40 USD. Each direction is half a spread away from the original midpoint; the round trip includes both halves. This assumes the quotes did not change and enough quantity was available for both trades. A real market may move before the second order arrives. [1] [3]

Now change the available quantities. Only 0.1 BTC is offered at the 60,010 USD ask; another 0.9 BTC costs 60,030 USD per bitcoin. If Ela buys a full 1 BTC across both levels, the quantities actually taken determine her bill: the average is 60,028 USD per bitcoin. That is 18 USD above the initial ask and 28 USD above the 60,000 USD midpoint. A narrow spread at the top of the book does not by itself make a large order cheap to execute. Market depth means the quantity available at the different prices. This calculation is our model, not an exchange’s fee formula. [2] [3]

Statistics may also show an effective spread, based on a completed trade rather than displayed offers alone. In a separate example, Ela buys at 60,006 USD against a reference midpoint of 60,000 USD. The one-sided difference is 6 USD. A convention described by the SEC, the U.S. securities regulator, for measuring stock executions doubles that difference to 12 USD. It measures execution quality; it is not another 12 USD fee. The time at which the methodology takes its reference quote matters too. This example explains a stock-market metric, not legal rules that apply to every bitcoin exchange. [4]

Return to the purchase at the original ask. For 1 BTC, the price is 10 USD above the chosen midpoint. If the provider additionally charges an explicit fee of 5 USD, the cost relative to that midpoint is 15 USD. We account separately for the actual purchase price’s deviation and the separately charged fee. Do not add the full 20 USD spread again to a deviation that already includes the difference from the midpoint. That would count part of the same cost twice. These amounts are illustrative, not current fees; identify any further costs separately. [3]

A comparison site may show Ela a high bid on one exchange and a low ask on another. They are not the two sides of one order book. First she needs to compare the same instrument and currency, contemporaneous prices and available quantities. Only then can she consider whether she can access both trades, has the necessary funds in place and faces fees or transfers. Numbers may look attractive on screen while simultaneous execution remains unavailable. Crossing prices alone therefore do not establish a guaranteed arbitrage: a profit from the price difference that can certainly be executed. [2] [3]

While reading statistics, Ela may encounter a missing value. CME’s methodology, for example, does not calculate the spread at a level when the bid or ask is absent. Blank is not zero: it does not say she can currently buy and sell at the same price. Similarly, a daily average summarizes observations over time; it is not an offer waiting for her order. Before comparing, she checks whether she is looking at live quotes, a time average or a price for a specified quantity. That makes the spread useful information rather than an isolated number. [2]