Library · Crypto, word by word · Trading and derivatives
What is slippage tolerance?
Slippage tolerance is a setting on a decentralized exchange swap that says how far the final price may move against you, as a percentage, before the trade is cancelled. If the price moves more than that before the transaction is processed, the swap fails instead of executing.
What it means for you
Set high, slippage tolerance lets bots sandwich your swap and fill you at the worst price you allowed. Set too low, swaps fail in busy or thin markets, and on many chains a failed transaction still costs a network fee. The tolerance is the most you agree to lose to price movement, not an estimate of what you will lose.
A common mistake: “A higher slippage tolerance gets my swap a better price.”
In fact: It only makes the swap more likely to go through. The tolerance is the worst price you accept, and an attacker can push your fill right up to that limit.
How it works
On an automated market maker, a larger trade moves the price more: that is expected slippage. Between submitting a swap and its inclusion in a block, other trades can move the pool, causing unexpected slippage. So the swap carries slippage protection and reverts if the execution price is worse than the limit. A sandwich attacker trades just before and just after the victim, pushing the price as far as the victim's protection allows, so a wider tolerance permits a larger attacker profit. Very low or zero tolerance makes trades likely to fail under heavy volume.
An example
Say a swap quotes 1,000 tokens and your tolerance is 1 percent, so the minimum you accept is 990. If the pool moves and you would get 985, the swap fails. If you would get 992, it goes through, even if a bot caused the drop.
Source: Zhou, Qin, Ferreira Torres, Le and Gervais: High-Frequency Trading on Decentralized On-Chain Exchanges (arXiv) · checked 4 October 2026
Often confused with
Related words
Educational content, not financial advice. Written by hand and checked against the source named above. Something wrong? Tell us and we reply within two business days.