Set tolerance to the largest adverse execution move you would still accept, then keep it tight enough to reject a materially worse fill. The right value depends on pool depth, trade size, volatility during confirmation, and route length; a volatile token’s label alone does not justify a wide threshold.
- Tolerance sets the minimum output, not the trade’s expected price impact.
- Use the quoted output and your acceptable loss to calculate a concrete minimum.
- If the quote needs a much wider threshold to execute, reassess the route or timing.
What does slippage tolerance actually protect?
For an exact-input swap, the router typically encodes a minimum output as quoted output × (1 − tolerance). If execution would deliver less, the transaction reverts; if it delivers more, you receive the better amount. A 1% setting on a quote of 10,000 output tokens therefore sets a minimum of 9,900.
This threshold is separate from price impact. Price impact is the pool movement caused by your trade against current liquidity; execution slippage is the additional difference between the quoted state and the state when your transaction is executed. In a constant-product pool, for example, output is approximately (1 − fee) × input × output-reserve / (input-reserve + (1 − fee) × input). A shallow reserve or large order can make the quote itself poor before tolerance enters the calculation.
For the full trade flow, see how Blackhole swap prices an Avalanche trade. When using Blackhole swap, treat the displayed quote as a snapshot: other swaps can move the pool before your transaction lands, and a multi-hop route can expose you to movement at more than one pool.
How wide should the threshold be?
Start with the trade’s acceptable shortfall, then compare it with the route’s likely movement during confirmation. As a working range, liquid pairs in ordinary conditions may need roughly 0.1–0.5%; volatile but reasonably deep pairs often call for 0.5–1%. Thin pools or fast markets can require 1–3% or more, but that is a sign to scrutinize the quote, not a default recommendation.
Example: before adjustment, a quote is 10,000 tokens and the tolerance is 0.5%, so minimum output is 9,950. If the pool moves 0.7% against the route before inclusion, the swap reverts and costs gas. After raising tolerance to 1.5%, minimum output becomes 9,850: execution can succeed at that worse rate, but you have accepted up to 100 fewer tokens than the original quote. Raising the setting buys execution probability by increasing adverse-price exposure.
On concentrated-liquidity routes, a price move can cross a tick into a region with different active liquidity, so a small spot change may produce a larger output change than recent average volatility suggests. Multi-hop routes compound pool fees and state changes, although the router’s final minimum usually protects the total route rather than enforcing a separate tolerance at every hop. A fee-on-transfer or otherwise nonstandard token can also make quoted and received amounts diverge or cause a revert; widening tolerance is not a reliable fix for incompatible token behavior.
How do you choose and apply it quickly?
Use a short repeatable check before submitting frequent trades. Blackhole swap is one way to make this kind of Avalanche exchange; the same minimum-output logic applies across decentralized exchanges and routes.
- Refresh the quote immediately before submission and note its output, route, and estimated price impact.
- Decide the largest output shortfall you accept in token or percentage terms; calculate the corresponding minimum output.
- Set tolerance near the movement needed to clear that minimum, using a lower starting point for deep pools and a higher one only when volatility or thin liquidity supports it.
- Check that the transaction deadline is short enough to reject a stale quote, then submit without waiting if the market is moving.
- If it reverts, inspect whether the cause was a changed quote, a tick crossing, route liquidity, or token behavior before retrying; refresh first instead of repeatedly widening the threshold.
For repeated trades, record revert frequency and realized shortfall by pair and route, then tune from observed execution rather than a universal preset. I’d accept a failed swap and its gas cost when the alternative is a minimum output I would regret; tolerance should encode that decision explicitly.