Chapter 06 · The race
Same payment.
Two networks.
Press pay. Kaspa settles in 1 to 3 seconds; Bitcoin needs about an hour. Each green square is a real Kaspa block mined after you pressed; Bitcoin’s six orange slots fill only as real Bitcoin blocks arrive.
Kaspa
0.0 s
Ready.
Bitcoin
0:00
Ready.
See both clocks on one ruler
What if most miners refuse your payment?
Slide how many miners refuse you, and watch Bitcoin’s wait and Kaspa’s wait side by side.
Speed is really about fairness. A censoring miner can only delay you until an honest one makes a block. Bitcoin gets a new block maker every 10 minutes, Kaspa ten a second. Same odds, very different wait.
One roll of the blocks above: the wait on this run. The average wait is in the caption under the figure.
The clock assumes ten blocks a second, the rate you watched live at the top of the page. Next, Who got the coins follows who gets paid for all those blocks.
When the internet breaks
GHOSTDAG fixes its safety margin k in advance, from an assumed worst-case delay. If the real network is much slower, a fixed margin can be too tight. , Kaspa’s planned next consensus, measures the network as it goes: fast when the internet is calm, slower but still safe when it splits. Core developers merged more DAGKnight code toward testnet in September 2026; it isn’t on mainnet.
The same live blocks also run as a street of buses in the home page live panel, Street tab: the second door to Kaspa Street.
Ten times the blocks, same coin flow
Block time 1,000 ms → 100 ms. The reward is set per second and split across the blocks. Live since May 5, 2025, at DAA score 110,165,000. Source: KIP-14.
The one sentenceA Kaspa payment is as safe in 1 to 3 seconds as a Bitcoin payment is after about an hour.
The exact version
The same chapter with the workings shown: the derivations, the edge cases, and the numbers behind the plain version.
Sources and the raw numbers
Where “settled” comes from
Kaspa side: “settled” uses Yonatan Sompolinsky’s figure of 1 to 3 seconds for the guarantee Bitcoin gives after about an hour. Bitcoin side: six confirmations, the usual rule for large amounts.
Where “1 to 3 seconds” comes from
For Kaspa, Sompolinsky gives figures on X (Feb 26, 2026). A miner with 37% of all mining power has a 12% chance of passing itself off as the majority at 10 blocks a second, and 0.3% at 100. That majority check is what decides censorship and ordering, which is why the core developers want more blocks a second, not more transactions. He also writes that Bitcoin reaches censorship resistance in about an hour and Kaspa reaches the same level in 1 to 3 seconds. Post, post. That’s why core wants more blocks per second, not for throughput. “Scalability is boring,” in his words.
More depth: Read the deep version: Seconds versus an hour