Every miner (the small glowing nodes in the ring) races to find the next block by repeated random guessing — proof-of-work. The network's overall find-rate is set by Difficulty: on average one block appears somewhere every N seconds regardless of how many miners are competing for it. Which miner wins a given block is decided by a weighted lottery: a miner's chance is proportional to its share of total hashpower, controlled here by Hashpower concentration.
P(block found this tick) = 1 − e^(−dt / difficulty)
P(miner i wins) = hashShare(i) / Σ hashShare
Network latency models the time it takes a freshly mined block to reach every other miner. If a second miner finds a competing block before hearing about the first one, both blocks exist at the same height — a fork. The chain then temporarily splits into two candidate branches (drawn one above the other). Whichever branch other miners happen to extend next grows longer; the instant one branch pulls ahead, the shorter branch is abandoned — its blocks turn grey and are removed as orphans, and every miner resumes work on the single longest chain. This is the longest-chain consensus rule: no central authority ever decides which branch is "true", the network simply converges on whichever accumulated the most proof-of-work.
- Miners — how many independent nodes are competing; more miners racing at once make simultaneous discoveries (forks) more likely.
- Difficulty — target seconds between blocks network-wide; lower difficulty means faster blocks but also a shorter window to avoid collisions.
- Network latency — how slow block announcements propagate; higher latency directly raises the fork rate.
- Hashpower concentration — how lopsided mining power is; pushed high enough, one miner approaches the 51% threshold where it could, in principle, out-race the rest of the network on its own — the classic centralization risk.