
A swap rarely costs what the headline rate suggests. Five things eat the difference: price movement while you wait, network fees on both legs, the hour you trade, thin liquidity on the pair, and routing through an intermediate coin. Below: each pattern with figures, plus a worksheet for calculating your real cost.
On Fswap the quote appears before you send and a fixed rate locks it, which removes the single largest variable below — 4,300+ crypto pairs, no account, on the web or in a Telegram mini app.
Sandwich attacks cost victims roughly 0.05%–0.5% per trade on volatile pairs, but 1%–5% on larger trades or illiquid pools without protected routing (Skrumble, May 20, 2026).
MEV extraction across all chains has passed $7.2 billion since 2020, running $10–20 million a day normally and $40–50 million during volatility spikes; sandwich attacks alone account for about $2.2 billion (Skrumble, May 20, 2026).
Liquidity is not constant: BTC/USDT and ETH/USDT spreads run 15%–30% tighter during 13:00–17:00 UTC than overnight, while 02:00–06:00 and 21:00–23:00 UTC are the thinnest windows (FinanceFeeds, Jul 7, 2026, updated Sep 7, 2026).
Network fees differ by two orders of magnitude between chains: sending USDT costs under $0.001 on Solana, around $0.10 on BNB Chain and 2–10 on Ethereum (Eco, 2026).
Bitcoin's own fee is currently small — the average transaction cost $0.2805 on October 5, 2026 — so on BTC legs the rate movement during confirmations matters far more than the fee (YCharts, Oct 5, 2026).
Every one of these shows up as the same complaint — "I received less than the calculator said" — and each has a different cause and a different fix.
Pattern | Typical cost | Worst case | The fix |
|---|---|---|---|
Thin pair, price impact and sandwich exposure | 0.05%–0.3% on small retail swaps | 1%–5% per trade on illiquid pools or large amounts | Trade liquid pairs; split large amounts; avoid unprotected DEX routing |
Double swap through an intermediate coin | Two spreads plus two network fees | Roughly doubles both cost components | Use a direct pair where one exists |
Rate drift on a floating quote | Fractions of a percent over 5–15 minutes | Several percent if the deposit is slow and the market moves | Take a fixed rate and send immediately |
Network fees on both legs | Under $0.01 to about 1oncheapchains | 2–$10 per Ethereum leg, regardless of the amount | Choose the network, not just the coin |
Trading in a thin hour | Spreads 15%–30% wider than peak | Compounds with pattern 1 on weekends | Prefer 13:00–17:00 UTC for large amounts |
Ordered by the largest realistic loss each pattern can cause, worst first. Figures from Skrumble (May 20, 2026), FinanceFeeds (Jul 7, 2026) and Eco's 2026 network comparison. Percentages are ranges reported for typical conditions, not guarantees.
A floating quote is an estimate until the moment your deposit confirms. Between clicking and confirming, the market keeps moving — and that gap is minutes, not seconds.
Bitcoin's average hourly volatility is roughly eleven times a dollar stablecoin's, and it is not evenly spread across the day: volatility peaks around 13:00 UTC, when the US market opens, with a second spike near 00:00 UTC, and bottoms out between 03:00 and 09:00 UTC. A swap that takes 5–15 minutes in normal conditions can therefore land anywhere inside a meaningful price range if you picked a floating rate at a busy moment.
Worked example. You swap $2,000 worth of BTC on a floating quote. The market moves 0.4% against you in the twelve minutes your deposit takes to confirm — unremarkable at 13:00 UTC. You receive about $8 less than the calculator showed. On $20,000, the same move is $80.
The fix: a fixed rate transfers that risk to the service, and you send immediately rather than when you get back to your desk.
Every swap is two transactions: yours in, theirs out. Both pay a network fee, and neither scales with your trade size — which is why small swaps hurt proportionally more.
The spread between chains is enormous. Sending USDT costs under $0.001 on Solana, about $0.001 on Polygon, roughly $0.10 on BNB Chain, and 2–10 on Ethereum. On TRON the same transfer is free to about $1 when the wallet has energy, but 2–4 when TRX is burned instead — and a first transfer into an empty wallet needs roughly 130,000 energy against 65,000 for one that already holds USDT. Bitcoin, by contrast, is cheap right now: the network average was $0.2805 per transaction on October 5, 2026.
Worked example. A $100 swap into USDT on Ethereum can lose 2–10 to the payout fee alone, or 2%–10% of the trade. The same swap into USDT on TRON with energy available costs closer to nothing. Our breakdown of gas fees compared across networks has the per-chain detail.
The fix: pick the network deliberately. The coin is the decision people agonise over; the network is the one that costs money.
Liquidity has a timetable. Aggregate volume across major exchanges peaks between 13:00 and 17:00 UTC, when European afternoon overlaps North American morning. The thinnest windows are 02:00–06:00 and 21:00–23:00 UTC.
That shows up directly in the price you get: BTC/USDT and ETH/USDT spreads run 15%–30% tighter at peak hours than overnight. Weekends are worse — the withdrawal of institutional desks and market makers removes roughly 65% of typical weekday Bitcoin volume, while about 35% of weekly transactions still happen then. Institutions account for around 63% of Bitcoin volume during US business hours, and when they are away the order book thins out underneath you.
There is a genuine trade-off rather than a simple rule: peak hours bring the tightest spreads and the highest volatility at the same time. For a large amount, tighter spreads matter more; for a floating quote, volatility does.
The fix: route large swaps through 13:00–17:00 UTC on a weekday, and if you must trade at 04:00 UTC on a Sunday, use a fixed rate so the thin book is the service's problem rather than yours.
On an illiquid pair, your own order moves the price. The deeper you have to reach into the book, the worse the average price you get — and on public blockchains, a visible pending trade with loose slippage tolerance also invites sandwich bots.
The measured cost is uncomfortably specific. Sandwich attacks take roughly 0.05%–0.5% per trade on volatile pairs, and 0.05%–0.3% on retail swaps under $1,000 with moderate slippage settings. On larger trades or illiquid pools without protected routing, that rises to 1%–5% per trade. Across all chains, MEV extraction has exceeded $7.2 billion since 2020, with sandwich attacks alone around $2.2 billion — about 30% of Ethereum's MEV. Daily extraction runs $10–20 million on quiet days and $40–50 million during volatility events.
Worked example. A $10,000 swap on a thin pair that loses 2% to price impact and sandwiching costs $200 — forty times the $5 you might have saved by shopping for a marginally better headline rate.
The fix: trade pairs that actually have depth, check whether a direct route exists before accepting a two-hop one, and split genuinely large amounts into smaller transactions. The list of popular exchange pairs is a reasonable proxy for where the depth is.
When no direct pair exists, the route becomes A → B → C. Everything in this article then happens twice: two spreads, two network fees, two windows during which the rate can drift.
Worked example. Converting an altcoin to XMR via BTC means an altcoin → BTC leg and a BTC → XMR leg. Each pays its own network fee, each carries its own spread, and the BTC middle leg adds the slowest confirmation step in the chain. A direct altcoin → XMR pair, where one exists, removes an entire set of those costs.
The fix: check for a direct pair first. If the service only offers the two-hop route, price it as two swaps rather than one when you compare it against alternatives.
The honest comparison is not between headline rates. It is between the amount that lands in your wallet and the mid-market value of what you sent. Work through these five lines before you commit.
Cost component | How to measure it | Example: $1,000 BTC → USDT (TRC-20) |
|---|---|---|
Quoted rate vs mid-market | Compare the quote against a reference price from an independent source | On a fixed-rate quote, the figure you see is the figure you get — Fswap shows both fixed and floating before you send |
Deposit network fee | Your wallet's fee estimate for the sending chain | About 0.28onBitcoinattheOctober5,2026average |
Payoutnetworkfee | Whethertheservicedeductsitandonwhichchain | 0–$1 on TRON with energy; 2–4 if TRX is burned |
Rate drift risk | Expected minutes to settle × the pair's typical hourly move | Zero on a fixed rate; roughly 4–8 on a floating quote in a 0.4% move |
Price impact and routing | Whether the pair is direct and deep, or thin and two-hop | Negligible on BTC → USDT; 1%–5% on an illiquid pair |
Components listed in the order you encounter them, from quote to settlement. Fee figures from YCharts (Oct 5, 2026), TronSave (updated Jul 21, 2026) and Eco's 2026 comparison; the drift and impact ranges are worked examples using the percentages cited earlier in this article, not measured outcomes.
Write down the amount the calculator promises, then compare it to a reference price elsewhere.
Confirm which chain the payout uses, and what a transfer on that chain costs.
Decide fixed or floating — and if floating, estimate how long settlement will take.
Check whether the route is direct or passes through an intermediate coin.
For amounts above a few thousand dollars, check the clock: weekday, 13:00–17:00 UTC if you can.
Save the order ID and both transaction hashes so you can reconcile what actually arrived.
A fixed rate carries a slightly wider margin, because the service absorbs the price movement while your deposit confirms. People read that margin as the expensive option and pick floating — which is right in some cases and costly in others.
A fixed rate wins when any of these is true:
The exact amount matters. Paying an invoice or a supplier means arriving short is a problem, not an inconvenience.
The sending chain is slow. Bitcoin confirmations take minutes; minutes are long enough for the market to move against you.
You are trading in a volatile window. Around 13:00 UTC, or during a news event, the expected drift can exceed the fixed-rate margin.
The amount is large. A fraction of a percent on $20,000 outweighs the margin difference several times over.
Floating is reasonable for small amounts, on fast chains, in calm markets, when you are sending immediately and a few cents of drift is irrelevant.
Either way, the mechanics are worth knowing before you commit funds: Fswap quotes the rate before anything is sent and typically settles in 5–15 minutes, with network congestion as the usual cause of delay — the full sequence is on the how it works page.
Usually one of five reasons: the rate drifted while your deposit confirmed, network fees were charged on both legs, the pair was thin and your order moved the price, you traded in a low-liquidity hour, or the route went through an intermediate coin and paid every cost twice.
The difference between the price you expected and the price you got, caused by the market moving or your order consuming available liquidity. On public blockchains, a loose slippage tolerance also exposes a pending trade to sandwich bots, which cost victims about 0.05%–0.5% per trade on volatile pairs.
Roughly 0.05%–0.3% per trade for retail swaps under $1,000 with moderate slippage settings, and 1%–5% on larger trades or illiquid pools without protected routing. Lifetime sandwich extraction is around $2.2 billion, about 30% of Ethereum's MEV.
For large amounts, 13:00–17:00 UTC on a weekday, when volume peaks and BTC/USDT and ETH/USDT spreads are 15%–30% tighter than overnight. Avoid 02:00–06:00 and 21:00–23:00 UTC, and weekends, when roughly 65% of weekday Bitcoin volume is absent.
Solana and Polygon cost under a cent per transfer, BNB Chain about $0.10, and TRON is free to roughly $1 when the wallet has energy. Ethereum runs 2–10 per transfer, which can exceed 2% of a small swap on its own.
Floating is usually cheaper on paper because its margin is narrower. Fixed is cheaper in practice whenever the expected rate drift during settlement exceeds that margin — slow chains, volatile hours, large amounts, or any payment where the exact figure matters.
On thin pairs, often yes: smaller orders reach less deeply into the order book and reduce price impact. The trade-off is one network fee per transaction, so on expensive chains the saving can disappear into fees.
Stop comparing headline rates and start comparing what lands in your wallet. Pick the network before the coin, take a fixed rate whenever the settlement window is long or the amount is large, prefer direct pairs over two-hop routes, and move serious amounts during weekday peak hours. The five patterns above are predictable, which means all five are avoidable.
Want to see the real number before you commit? Get a live quote on Fswap — 4,300+ pairs, fixed or floating rate shown up front, no account, on the web or in Telegram.
Elena leads marketing at Fswap and writes about cryptocurrency exchange: how swaps work, how to compare rates and fees, and how to keep funds safe without unnecessary KYC.