Slippage is the difference between a trade’s actual execution price and the price used for comparison. First identify that reference price and its time. In this explanation, a positive price cost means a worse outcome for the trader and a negative one means a better outcome; separately charged fees are added separately.
Tomáš is about to buy BTC and records a price for comparison. What is he looking at? The last completed trade, the ask or best seller’s offer, the bid or best buyer’s offer, and the midpoint between them are different references. Coinbase’s rules also describe slippage against the last trade, whose price is no promise for the next one. To make the comparison meaningful, Tomáš records the venue, currency, time, direction and requested quantity too. Otherwise, people may disagree about a deviation simply because they measure it from different starting points. [1]
In our model, Tomáš buys 1 BTC. Only 0.4 BTC is available at 60,000 USD; he buys the remaining 0.6 BTC at 60,100 USD per bitcoin. Weight each price by the quantity purchased: the larger part of the trade contributes more to the result. The average is 60,060 USD per bitcoin. Compared with the reference of 60,000 USD chosen beforehand, this whole purchase has a price cost of 60 USD before fees. That is 10 basis points, or 0.1% of the reference value. A basis point is one hundredth of a percentage point. These are our illustrative figures, not current exchange quotes. [1] [3]
Reverse the situation: Tomáš sells 1 BTC. He sells 0.4 BTC at 60,000 USD and another 0.6 BTC at 59,900 USD per bitcoin. The weighted average is 59,940 USD. Against a reference of 60,000 USD, the cost is again 60 USD because he received less than the comparison price implies. For a purchase, subtract the reference from execution; for a sale, subtract execution from the reference. Simply subtracting the reference in both cases would give the sale the opposite sign and could make a worse outcome look like an improvement. [1] [3]
Consider a different purchase: Tomáš gets 1 BTC for 59,980 USD against a reference of 60,000 USD chosen beforehand. He saves 20 USD. Under our convention, the price cost is minus 20 USD: a price improvement. Slippage need not mean only a loss; check how the particular report defines it. Keep the meaning of the signs consistent across trades, and do not confuse a better execution price with a reduction in a separately charged commission. [2]
Tomáš knows the outcome, but not necessarily why it happened. His order may have consumed cheaper offers and continued through higher levels of the book, the list of waiting orders. The market may also have moved between his decision and execution. From the deviation alone, we cannot separate his own price impact from changes caused by others. Timestamped offers, available quantities and individual fills help assess what happened; a daily average cannot replace that sequence. Large slippage is a reason to examine execution, not by itself proof of exchange misconduct. [1] [3]
To set the highest acceptable buying price, Tomáš can use a limit order. It may buy at the limit or below; a sell limit permits selling at the limit or above. The order can remain unfilled or only partly filled. Its rules and duration determine whether the remainder waits or is canceled. A limit also protects its own chosen boundary, not every other benchmark Tomáš might later select. It gives him control over the acceptable price without promising the entire requested quantity on time. [1] [2]
Return to the first purchase. A price cost of 60 USD plus a separate fee of 12 USD gives 72 USD against the same reference and for the same quantity. These are illustrations, not actual fees. If the measured deviation already includes crossing the gap between buy and sell offers, do not add that spread again: part of the same cost would be counted twice. Likewise, a slippage-tolerance setting is a boundary for acceptable deviation under that service’s rules, not an amount automatically charged as a fee. [1]
Tomáš wanted 1 BTC, but only 0.4 BTC filled for a total of 24,000 USD. To find the average price, divide execution value by 0.4, not 1; the result is 60,000 USD per bitcoin. The remaining 0.6 BTC was not bought and must be reported separately. Spreading the expense across the entire intended purchase would produce an apparently attractive price nobody offered. Good-looking slippage on the filled portion can therefore accompany an unfinished plan. Assess the actual quantity and what happened to the remainder alongside the price. [1] [3]
For the clearest picture, read this entry together with Bid–ask spread, Market order, Limit order, Order book. The reverse links also lead from Market maker, Bid–ask spread, Market order, Limit order.