What Is an Orphan Block?

In everyday usage an orphan block is a valid Bitcoin block that nonetheless doesn't stay in the blockchain. Two miners found a block almost at the same time, the network briefly split into two camps, and the next block decided which side keeps growing. The losing block is more precisely called a stale block. Its miner gets nothing; the transactions in it end up in the winner's block or back in the mempool. As of November 2026 this happens only rarely, because blocks now travel through the network in fractions of a second.

Orphan block vs stale block: the difference

The terms get mixed up even in technical writing. The exact distinction still helps you understand what is happening.

Stale block: A fully valid block whose parent is known. It simply lost the race: another block at the same height became the basis of the longer chain. When people say "orphan block", they almost always mean this.

Orphan block in the strict sense: A block whose parent a node doesn't have yet. The node can't place it and has to fetch the parents first. Earlier versions of Bitcoin Core kept such blocks in a separate buffer. Since version 0.10 in 2015 a node downloads the block headers first and then the blocks in the right order, so true orphans hardly ever occur any more.

A picture: two runners cross the finish line almost together. The stale block is the narrow runner-up who ran just as fairly. The strict orphan is a runner the finish line doesn't yet know which race they started in.

What is in a block and how the reference to the previous block works is explained in What is a Bitcoin block?.

How a stale block comes about

A miner who finds a valid block sends it to its neighbours in the network immediately. They check it and pass it on. It takes some time until the block has reached all nodes and, above all, all other miners.

During that time the other miners keep working on the old state. If one of them also finds a block in exactly that window, there are two valid blocks at the same height:

1. The network splits. Each node builds on the block it saw first. Some miners keep working on block A, some on block B.
2. The next block decides. Whoever finds the following block first extends their side. That chain now has more work behind it.
3. Everyone switches. Nodes always follow the chain with the most work. Nodes on the other side switch over: they roll back the losing block and adopt the longer chain. This is called a reorg (reorganisation).
4. The loser becomes stale. Its reward doesn't exist in the valid chain. Transactions that were only in it go back to the mempool and wait for the next block.

This rule of always following the chain with the most work is the core of proof of work. It makes the network agree again within minutes, without a vote and without a boss.

How often it happens: from 2 % to almost zero

The rate has changed a great deal over Bitcoin's history.

2013: Christian Decker and Roger Wattenhofer of ETH Zurich measured how quickly blocks travel through the network. On average a block took 12.6 seconds to reach most nodes. Forks occurred at around 1.69 % of all blocks.

2016: A study led by Arthur Gervais, also at ETH Zurich, found a stale rate of around 0.41 % for Bitcoin.

Since then: Two technical changes cut propagation time sharply. With compact blocks (BIP 152, from Bitcoin Core 0.13 in 2016) a node no longer sends the whole block, only short identifiers of the transactions. The recipient already has most of them in its mempool and can rebuild the block itself. And large miners are connected directly via fast relay networks. A block thus reaches the other miners in fractions of a second.

As of November 2026 stale blocks are rare. Exactly how rare nobody knows for sure: a stale block seen by only a few nodes shows up in no statistic. Services such as ForkMonitor and datasets collected by developers count what they find and, by their own account, don't capture everything.

The newest blocks and who found them:

Longer forks: reorgs of two blocks and more

Usually everything is settled after one block. Occasionally both sides also find the next block, and the race lasts longer.

On 23.03.2026 there was such a case at height 941’880. AntPool and ViaBTC had each found a block, while Foundry found several in a row on the other side. Foundry's chain won, and two blocks from AntPool and ViaBTC went stale. In the discussion that followed, developer Anthony Towns estimated that such two-block reorgs happen roughly every 150’000 blocks, about every three years. Earlier cases occurred in 2013, 2015, 2020 and 2023.

Two historic exceptions were not races but bugs:

• August 2010: A software bug allowed a transaction that created around 184 billion bitcoin. Satoshi Nakamoto and others released a fixed version within hours. The corrected chain overtook the faulty one after around 50 blocks.
• March 2013: Version 0.8 of Bitcoin Core accepted a large block that older versions rejected because of a database limit. The network split for around two dozen blocks until the large pools, at the developers' request, went back to the old version.

Both cases show that a fork can also be caused by software. How rule changes are introduced today without splitting the network is described in Bitcoin soft fork explained.

Why stale blocks matter to miners

For a miner a stale block is a total loss. The electricity has been used, the reward doesn't exist. On 13.11.2026 a block was worth 3.149 BTC on average (3.125 new bitcoin plus fees), around €196’500 at a price of around €62’400.

That is why a miner may only spend the new bitcoin from a block after 100 further blocks. This waiting period ensures that coins which no longer exist after a reorg are never in circulation. More in the section on the coinbase transaction in What is a Bitcoin block?.

Size protects: the advantage of large pools

A large pool has a built-in advantage in the race. It doesn't compete with itself: its own miners immediately continue on its new block, with no propagation delay. The larger its share of hashrate, the smaller the risk that someone else strikes at the same moment.

This is one reason developers put so much effort into fast block propagation: high stale rates reward size and encourage concentration in a few pools, whose role is described in Mining pools.

Selfish mining: playing the race on purpose

In 2013 Ittay Eyal and Emin Gün Sirer of Cornell University described a strategy called selfish mining. A miner withholds a block it found and secretly works on the next one. When someone else publishes a block, it releases its own and, with a good connection, wins the race. That way it deliberately makes other miners' blocks stale and raises its own share of the rewards.

By their calculation this pays from a share well below 50 % of hashrate if the attacker wins races often enough. In practice no sustained selfish mining has been proven. But with reorgs like the one in March 2026, people debate every time whether it was chance, poor propagation or intent.

What stale blocks mean for your confirmations

As a bitcoin holder, one thing matters most to you: when is a transaction safe?

One confirmation can disappear. If your transaction is in a block that later goes stale, it loses its confirmation. As a rule it is also in the competing block or gets into the next one. It isn't lost; it just waits briefly.

It only becomes dangerous with fraud. A sender could try to pay you and at the same time send the same money to themselves on the other side of the fork. That only works with a great deal of hashrate or a lot of luck.

Each additional block makes it less likely. Satoshi Nakamoto calculated in the white paper how an attacker's chance shrinks with every confirmation.

How transactions get into the mempool and from there into a block is explained in The mempool.

What remains: a sign of healthy propagation

Stale blocks aren't a flaw but a consequence of thousands of miners worldwide working without coordination. The network resolves every fork by fixed rules and without intervention.

A low stale rate shows that blocks are distributed quickly and evenly. If it rises, that is a warning sign: blocks too large, connections too slow, or a miner playing games. For you as a saver, an occasional stale block changes nothing. After a few confirmations your coins are as firmly anchored in the chain as any other transaction.