Zcash block time simulator
Wallets and services pick their own depth.
What users would feel
What it would cost
Why the math is trustworthy
Share of block gaps lasting 2 minutes or more. The same random-arrival model matched every speed Zcash has run at, within about a percentage point.
| Era | Model | Measured |
|---|---|---|
| 150s Before Blossom | 45.1% | 44.7% |
| 75s Before NU7 | 20.3% | 20.7% |
| 25s With NU7 | 0.84% | 0.88% |
Compared with other chains
Protocol block targets. Chains settle differently, so only clear, documented waits are compared.
Block time (log scale)
Bitcoin10 min
Litecoin2.5 min
Monero2 min
Dogecoin1 min
Zcash · your setting25s
Cardano20s average
Ethereum12s slots- Kaspa0.1s (10 per second)
Usual wait until done
Bitcoin60.0 min6 confirmations, a common convention
Monero20.0 min10-block lock before funds can be spent
Litecoin15.0 min6 confirmations, a common convention
Ethereum12.8 minfinality after two epochs (64 slots), nominal
Zcash at 25s75s3 confirmations, your choice above
method_and_limitationsThe model, the measured data behind it, and what it cannot tell you.
Proof-of-work blocks arrive at random. With difficulty holding the average at the target T, gaps between blocks follow an exponential distribution: the share of gaps lasting at least x seconds is e−x/T. Waiting for k confirmations is the sum of k such gaps. Because arrivals are memoryless, a transaction waits a full average gap for its first block, not half of one.
This model was checked against canonical mainnet block-header times:
- 150s, before blossom (blocks 636,320–653,599, 17,280 gaps): average 150.68s; gaps of 2+ minutes 44.7% measured against 45.1% predicted.
- 75s, before nu7 (blocks 3,508,000–3,543,404, 35,405 gaps): average 75.36s; gaps of 2+ minutes 20.7% measured against 20.3% predicted.
- 25s, with nu7 (blocks 3,543,406–3,579,544, 36,139 gaps): average 25.13s; gaps of 2+ minutes 0.88% measured against 0.84% predicted.
The typical longest wait in a day is the expected maximum of a day's gaps, T × (ln n + 0.577) for n blocks. At slower targets the measured daily longest runs up to about 8% above it.
Orphans are blocks that lost a real race: a competing block that reached at least one of ZecBlock's observers no more than 2 seconds after that observer received the winning block, or that our own node adopted before switching away. Blocks arriving later everywhere are late blocks, usually a pool mining on an outdated tip; they waste work but never competed for the chain, so they are left out. Observers grew from 1 to 10 on 5 November 2026, so only the NU7 era has a comparable count: 225 orphans, 47 late blocks and 31,096 canonical blocks over the complete UTC days 7–15 November, 0.72%, about 25 a day. Live figures are on Mining. Between 10 and 25 seconds the page estimates the rate by scaling today's with block rate, assuming a fixed propagation delay; below 10 seconds orphans can rise much faster than that, so no estimate is shown. Observers do not see every competing block, so these are lower bounds.
Empty-block data multiplies blocks per day by the average size of a block holding only its coinbase transaction since NU7 (1,943 bytes). Confirmation depth is a choice: ZIP 315 recommends 10 for received funds in wallets, while many payments use fewer.
Other chains are listed at their protocol block targets. Their waits use each chain's own rule: a common confirmation convention for Bitcoin and Litecoin, Monero's 10-block lock, and Ethereum's nominal two-epoch finality, which often takes longer in practice. Proof-of-stake and blockDAG chains do not follow the random-arrival model, so they are shown for reference only.
This page shows consequences of a block time, not a recommendation. Changing it needs a network upgrade, and things the model does not cover, such as network bandwidth, mining pool behaviour and wallet sync, also matter. See measured block times on Zcash block speed.