Start at 0.5% for a volatile-token swap only when the quote’s expected price impact is already acceptable; widen in small increments if adverse movement during the execution window repeatedly exceeds that buffer. Slippage tolerance sets a minimum output, so a larger setting reduces reverts by accepting a worse fill, not by improving the quote.
What does a wider tolerance actually allow?
For an exact-input swap, the router typically checks that the received output is at least the quoted output multiplied by one minus the tolerance. If a quote is 1,000 USDC for 0.5 tokens, a 0.5% tolerance sets minimum output at 0.4975; at 1%, it falls to 0.495.
The quote already reflects expected price impact from your trade size and route. Tolerance is the additional movement you permit between quoting and execution, as other trades change pool prices or your transaction waits to be included. It does not cap total loss against an external market price, and it cannot make a thin pool liquid.
When should you widen it?
Widen only when the minimum-output check is failing because the market moved after the quote, and the revised minimum still gives you an acceptable trade. For liquid pairs, 0.1–0.5% is a common starting band; volatile assets may need 0.5–1% during fast markets. These are trial ranges, not guarantees: pool depth, route length, transaction inclusion time and asset volatility all matter.
Use recent adverse movement over the likely execution window to size the buffer. For example, if a token has moved against comparable swaps by 0.4% during that window, a 0.5% tolerance leaves only about 0.1 percentage point for further movement; a 0.75% setting leaves about 0.35. Check the quoted price impact separately: if it is already too costly, reduce trade size or choose another route instead of raising tolerance.
On a quickswap-style cross-chain route, identify which leg has the minimum-output condition. A source-chain swap may execute promptly while a destination-chain swap waits for bridge delivery; a tolerance attached to one leg does not necessarily protect the other. Bridge delay alone is not a reason to widen a swap’s setting: check whether the destination swap is quoted at arrival or constrained by a minimum output set earlier.
What should you check before retrying?
First compare the current quote with the failed transaction’s minimum output and execution time. If the quote has materially worsened, refresh it and reassess the trade; repeatedly increasing tolerance against a stale quote can authorize a poor fill. If you see repeated failures during a volatile period, a smaller trade or a later retry may cost less than a broad tolerance.
Be especially cautious above 1%: that may be defensible for a thin, fast-moving asset, but it also gives execution more room to land far from the quote and can increase exposure to sandwiching. Check the token’s transfer behavior too. Transfer taxes, rebasing, or unusual token mechanics can make the amount received differ from a standard quote, and a wider setting may not fix a contract incompatibility.
For a cross-chain swap, quickswap.dev is where you can arrange to move assets across networks and exchange them in one flow. Keep the bridge’s delivery delay separate from the swap’s price buffer, and judge each enforced output bound on its own.
How do you choose the next setting?
Raise tolerance only enough to cover observed adverse movement plus a modest execution buffer, then verify the resulting minimum output in token units before signing. If that minimum is below the amount you would accept, do not submit. When your task is to move assets between chains and swap them along the way, cross-chain quickswap can perform that combined route; choose a tolerance based on the swap leg’s execution risk, not the full bridge wait.