Slippage is the difference between the price you expected and the price you actually got. It happens because the market moved, or because your order was larger than the liquidity sitting at the best available price.
On deep markets it is negligible. On thin ones it routinely costs more than the trading fee, which is why comparing exchanges on advertised fees alone misses the larger number.
Why slippage happens
Three causes, which behave differently.
Order size against depth. Your order climbs the order book, filling at progressively worse prices. This is the most common cause and the most predictable.
Price movement between submission and execution. Markets move continuously. A few hundred milliseconds is enough during volatility.
Transaction ordering. On-chain, other trades can settle before yours, moving the price you land at. Bots exploit this deliberately.
A worked example of slippage
Two markets, same quoted price of $100, and you want to buy $50,000.
On the deep market, sell orders sit stacked at $100.00, $100.01, and $100.02, each holding hundreds of thousands of dollars. Your order fills almost entirely at $100.00. Average price: about $100.01. Slippage: one basis point.
On the thin market, there is $5,000 at $100.00, then $5,000 at $101.50, then $8,000 at $104.00. Your order climbs. Average price: roughly $106. Slippage: 6%.
Same headline price. Same trade size. One cost sixty times more, before any fee was charged.
Then consider the exit, where the identical problem runs in reverse.
Positive slippage exists too
Rarely mentioned, and real.
If the price moves in your favor between submission and execution, you fill better than quoted. A buy order placed just as the price ticks down gets more for the same money.
Most platforms pass this through. Some quietly do not, which is worth checking in the fee documentation.
Slippage tolerance and sandwich attacks
On decentralized exchanges you set this yourself, and the setting is a trap in both directions.
Set it too low and the transaction reverts, wasting the gas fee with nothing to show. Set it too high and you authorize the trade to execute at a much worse price, which is precisely what sandwich attacks exploit.
Here is how a sandwich attack works. A bot sees your pending transaction with a 5% tolerance. It buys ahead of you, pushing the price up. Your trade executes at the worse price, within your tolerance. The bot then sells into the price your buying created. Everything within tolerance is profit for someone else.
Setting tolerance close to expected market movement, rather than generously, is what limits the exposure.
Where slippage hurts most
- Small-cap tokens. Thin books mean a modest order moves the price several percent.
- Market orders during volatility. Speed is guaranteed. Price is not.
- Stop-loss orders triggering into a crash. The mechanism activates precisely when depth has evaporated.
- Off-hours trading. Liquidity thins outside the dominant regions' active periods.
- Large orders relative to daily volume. When your trade is a meaningful share of a day's activity, you are the market rather than a participant in it.
- Decentralized exchange pools. Pool size determines price impact. A $500,000 trade against a $1 million pool moves it dramatically.
How to reduce slippage
Descriptively, since the right approach depends on what you are trading.
- Check depth before size. Comparing your intended order against visible book depth is the single most useful check.
- Use limit orders where execution certainty is not the priority. You control the price, and you accept it may not fill.
- Split large orders across time or price levels rather than submitting one block.
- Trade the deepest pair. The same asset can be liquid against USDT and nearly untradeable against a less common quote currency.
- Avoid thin hours for large trades.
- Set tolerance tightly on-chain, and accept the occasional failed transaction as the cost of not being sandwiched.
What slippage is not
- Not a fee. Nobody charges it. It emerges from market structure.
- Not always negative. Favorable moves produce positive slippage.
- Not the same as the spread. The spread is the gap between best bid and best ask. Slippage is what happens when your order goes beyond that first level.
- Not fixed. The same order at the same size costs different amounts depending on when you submit it.
Put this into practice on mb.io
Slippage is a property of where you trade rather than something you can add protection against afterwards.
mb.io is a regulated crypto spot exchange backed by MultiBank Group, a financial institution founded in 2005 that serves more than 2 million clients across 100+ countries.
- Regulated by VARA in the UAE and AUSTRAC in Australia
- 40-nanosecond execution speed
- 10/10 security score from Hacken, an independent blockchain security auditor
- Institutional-grade MPC custody powered by Fireblocks, with segregated client funds
- A curated list of assets, so you're not trading into thin books on tokens nobody supports
- 24/7 multilingual customer support
Open your account and start trading on mb.io.
Frequently asked questions
What causes slippage in crypto?
Three things: your order being larger than the liquidity at the best price, the market moving between submission and execution, and other transactions settling ahead of yours.
Is slippage the same as a trading fee?
No. A fee is charged by the exchange. Slippage emerges from market structure and nobody collects it, though on thin markets it frequently costs more than the fee does.
What is a good slippage tolerance?
It depends on the asset's normal volatility and pool depth. Too low and transactions revert, wasting gas. Too high and you authorize execution at a much worse price, which sandwich bots exploit.
Can slippage be positive?
Yes. If the price moves in your favor between submission and execution, you fill better than quoted. Most platforms pass this through, though not all do.
How do I avoid slippage on large orders?
Check book depth against your intended size first, then consider splitting the order across time or price levels, and use limit orders where execution certainty is not the priority.
Why is slippage worse on decentralized exchanges?
Because price comes from pool reserves rather than an order book, and pool size directly determines price impact. Transaction ordering on-chain also lets bots sequence trades around yours.
What is a sandwich attack?
A bot buys ahead of your pending transaction to push the price up, lets your trade execute at the worse price within your tolerance, then sells into the move. Wide slippage tolerance is what makes it profitable.
Does slippage affect stop-loss orders?
Considerably. A stop triggers a new order into whatever depth exists at that moment, and crashes are exactly when depth disappears. The stop price sets when you enter the market, not what you receive.
