Skip to content
A global perspective on digital financeAbout The Token Press
CRYPTO. FINANCE. WHAT’S NEXT.The Token PressYour daily briefing ↗

Intermediate · 6 min read

Slippage vs Price Impact: How to Read a Crypto Swap Quote

Separate price impact, quote changes, minimum received and fees. Use a fictional swap worksheet to understand what a slippage limit does and does not protect.

The Token Press education desk · Updated

Part of our DeFi learning hub: What is DeFi.

What is the difference between slippage and price impact?

Price impact describes the effect of your trade on the available liquidity and resulting price. Slippage describes how the execution differs from the expected quote as conditions change. A slippage tolerance limits what a transaction may accept; it does not remove price impact already reflected in that quote.

Uniswap's help separates these concepts and explains why trade size relative to liquidity matters. A quote can already be poor before you choose a tolerance. Raising that tolerance gives execution more room to worsen; it does not make the pool deeper or restore the opening market price.

This guide is an offline quote-reading exercise, not a recommendation to trade or a universal tolerance setting. The existing liquidity-pool guide explains the underlying pool mechanics. Here we follow the numbers from a reference price to a quote, an acceptance boundary and a completed result.

Read the amounts before the percentage

An exact-input swap fixes the amount you sell and generally constrains the minimum output. An exact-output swap fixes the desired output and constrains the maximum input instead. Uniswap's quote API explicitly distinguishes these directions. Make sure you know which amount your interface allows to vary.

Expand the details and record the input and output tokens, network, amount, quote time, route, fees, network estimate and displayed execution boundary. Uniswap's swap-screen documentation identifies separate fields for fees, network costs, price impact and maximum slippage; the latter can expose the minimum output.

Read the actual minimum received or maximum paid value. Percentage labels, rounding and fee treatment depend on the implementation, so do not reconstruct a contract boundary from a remembered formula. A limit denominated in tokens also does not fix those tokens' future dollar value.

A quote involving another network, a wrapped asset or a different token contract is not an interchangeable price comparison. Identify the exact assets first. Use the token-contract guide to check identity without connecting a wallet.

Worked example: reference, quote, minimum and execution

Consider a fictional exact-input swap of 10 A for B on one network. At the starting reference rate of 100 B per A, simple multiplication suggests 1,000 B. The simulated route instead quotes 970 B because of the trade's effect on liquidity. For this first worksheet, assume no trading, token or interface fees, no other route differences, and unchanged market conditions while the quote is calculated.

The shortfall from the reference is 1,000 minus 970 = 30 B, or 3% of the 1,000 B reference. Under these deliberately isolated assumptions, that is our price-impact measure. A real interface may use a different benchmark or include other effects; a difference between two unrelated screenshots does not establish price impact.

Suppose the transaction's displayed minimum received is 960 B. We take that boundary as given rather than deriving it from a particular platform's tolerance algorithm. An eventual output of 962 B passes this output check; an output of 958 B should fail it in a correctly enforced, atomic pool swap. Other checks can still fail even when the amount is sufficient.

At 962 B, the additional shortfall against the 970 B quote is 8 B: 8 divided by 970 times 100 is about 0.825%. Against the original 1,000 B reference, the total shortfall is 38 B, or 3.8%. These percentages use different denominators. Simply adding 3% and 0.825% is not the correct total.

Changing the minimum from 960 B to 930 B does not raise the 970 B quote. It permits a worse output to pass that boundary. Even a successful transaction can therefore be a result the reader would not have wanted.

Visual explainerThree different numbers in one swap
Reference output1,000 BStarting rate multiplied by the input; not a fill promise.
Route quote970 BIncludes the assumed trade impact in this fee-free example.
Displayed minimum960 BOutput acceptance boundary; other checks still apply.
A 962 B result clears the output boundary but remains 38 B below the reference.Original hypothetical exact-input example. No real prices, live quotes or platform-specific tolerance formula.

Keep fees in a separate, unit-labeled worksheet

For a second fictional example, assume a route quotes 500 B and explicitly states that its pool fee is already included. A separate interface fee of 2 B is excluded from the quote. The expected recipient amount is then 498 B, before any separately paid network expense. Subtracting the included pool fee again would double-count it.

Assume network gas costs 0.001 G, where G is the native fee token. Do not subtract 0.001 from 498 B: these are different assets. If, only for this worksheet, G is valued at $2,000 and B at $1, the gas estimate is $2 and expected output less that separately funded cost is $496. These invented valuations are neither prices nor evidence of profitability; input acquisition costs are not included.

Real fee placement needs inspection. Uniswap's API, for example, documents an optional integration fee whose amount is reported separately from the quoted amount in specified cases. That is why an API output, an app preview and a final recipient balance need not mean the same thing.

Record which fees are included, which are deducted later, and which require an additional balance or approval transaction. The output boundary must be interpreted using the actual router and token behavior, not a generic worksheet assumption.

Why can a swap fail even with a wide tolerance?

A transaction can fail for reasons unrelated to normal price movement, including an expired deadline, an unmet approval requirement or incompatible token behavior. Inspect the failure reason and route instead of repeatedly increasing tolerance. Uniswap documents simulation and deadline checks as part of its pool-routing process.

Separate an on-chain reverted transaction from a preview that never submitted anything. On-chain execution can consume gas even when the swap reverts. A front-end error alone does not prove a gas payment occurred: inspect the transaction record. An unfilled UniswapX order is a different mechanism; its help page says no network cost is charged for that failed fill. A separate prior approval can still be its own transaction.

Transaction ordering is another concern. Ethereum's MEV documentation explains how a sandwich can place trades around a user's swap. A permissive boundary does not protect against every unfavorable ordering outcome. Use the route's documented protections and limits; no setting makes an unknown token or contract safe.

Stop if the asset, recipient, spender or requested permission differs from what you intended. A correct numerical limit is not a substitute for verifying what the transaction authorizes.

Five-minute practice: compare two complete quotes

Use two fictional quotes for the same input, assets and network at the same observation time. Route Cedar expects 1,000 B with a displayed minimum of 990 B and $6 of separately paid gas. Route Pine expects 998 B with a minimum of 988 B and $2 of gas. In this worksheet, B is valued at $1 and all other fees are already included.

Expected output less gas is $994 for Cedar and $996 for Pine. At their respective output boundaries, and assuming the stated gas expense, the figures would be $984 and $986. Pine's smaller headline output is larger after the assumed gas cost. These are scenario comparisons, not guaranteed net proceeds: gas estimates and token values can change.

Now change only Pine's gas estimate to $7. Its expected figure becomes $991 and its boundary figure $981. The output-token quote did not change, but the comparison did. Keep the original observation time and changed assumption in your notes rather than silently mixing old and new quotes.

Finally, ask whether the boundaries include every output deduction. If that field is unknown, neither worksheet establishes the actual minimum recipient amount. Resolve it in the provider documentation or leave the comparison incomplete. You can finish the exercise without signing a message or spending funds.

  • Compare the same input size, exact assets, network and observation time.
  • Distinguish reference price, quoted output and execution boundary.
  • Keep included fees, extra deductions and network gas separate.
  • Use a common valuation only when its assumptions are explicit.
  • Check the final receipt and recipient amount rather than calling the preview a completed trade.

Sources & further reading

Educational content. Examples are illustrative. Protocols, products and rules can change; consult the linked documentation for current details.

Explore the newsroom

Connect what you have learned with the latest reporting.