The Cheapest Times of Week to Transact, With Data
Network demand follows working hours in Europe and North America. The pattern is consistent enough to plan around.
Base fees are set by block fullness, block fullness is set by demand, and demand follows the working day of the regions where most activity originates. That produces a weekly pattern that has been stable for years.
The shape of the week
Weekday afternoons, European time, are the most expensive. This period overlaps European working hours with the North American morning. It is consistently the daily peak.
Weekday early mornings, European time, are cheap. North America is asleep and Europe has not started.
Weekends are cheaper throughout. Saturday and Sunday show materially lower base fees than any weekday, with Sunday typically the cheapest day.
Sunday early morning, European time, is the weekly low. This has been the cheapest window in every measurement period we have looked at.
The ratio between the weekly peak and the weekly trough is commonly three to five times. On an operation costing $20 at the peak, that is a difference of $14 to $15 for doing the same thing at a different hour.
Why the pattern is stable
It is not sentiment. It is the aggregate of automated systems, trading desks and users in two time zones, and none of those move.
The pattern breaks only when something unusual happens: a large mint, a liquidation cascade, a major protocol event. Those are unpredictable and short-lived, and they push fees far above anything the weekly cycle produces.
How to use it
For anything not time-sensitive, wait for the weekend. Consolidating wallets, revoking approvals, moving long-term holdings to cold storage. None of these are urgent.
Use your wallet’s slow tier. Most wallets offer speed options. The slow tier reduces the priority fee, not the base fee, so the saving is smaller than the timing effect but it is free.
Batch. Five transactions cost five fees regardless of timing. One transaction moving five times as much costs one.
Watch the base fee, not a headline. It changes block to block. A gas tracker gives you the current figure, which is the only one that matters.
Measuring it yourself
The method we use, for anyone who wants to verify rather than trust:
- Record the base fee from a block explorer at a fixed set of times each day for four weeks.
- Convert to a cost for a standard operation, using 21,000 gas for a transfer.
- Convert to local currency at the spot rate at that moment.
- Average by day of week and hour.
Four weeks is enough to see the weekly pattern clearly. Anything shorter is dominated by individual events.
The one methodological trap: convert to currency at the rate of the moment, not the rate today. A fee denominated in ETH looks different in dollars depending on when you price it, and mixing the two produces a chart that shows the ETH price rather than the fee.
What this does not apply to
Layer 2 networks. Their fees are dominated by a different mechanism and are low enough at all times that the daily variation is not worth planning around.
If you find yourself carefully timing mainnet transactions to save $12, the larger saving is usually to do the same operation on a rollup for cents. The timing strategy matters for operations that must happen on mainnet, which for most users is a small and shrinking set.
Fees move with network demand. Any figure here reflects the date shown above and should be re-checked before you act on it. Where a measurement was taken by hand, the article says so.
Related reports
- A Year of Fee Data: What We Learnedgas mechanics
- Batching: The Simplest Way to Cut On-Chain Costsgas mechanics
- Priority Fees, Block Building and Where Your Tip Goesgas mechanics