Glossary / Culture & history
Block size war
Also known as Blocksize war.
- Definition
- The block size war was the 2015 to 2017 fight over whether bitcoin should raise its block size limit by hard fork, and it ended with SegWit and a chain split.
At stake was who gets to change bitcoin. One camp wanted bigger blocks and cheaper payments; the other wanted a limit small enough that ordinary people could keep verifying the chain themselves. The dispute ran for roughly three years, produced at least four competing software releases, and settled the governance question far more decisively than the capacity question.
How it works
The block size war was fought over a single constant, and understanding why it could not simply be edited explains the whole conflict.
Raising the limit requires every participant to accept blocks that the old software rejects, which is a hard fork: nodes that do not upgrade stop following the chain. That turns a one line change into a referendum. Through 2015 and 2016 the big block side shipped its own implementations to try to win that referendum by adoption. Bitcoin XT, released in August 2015, carried Gavin Andresen's BIP101 schedule of 8 MB doubling every two years. Bitcoin Classic followed in early 2016 with a simpler jump to 2 MB, and Bitcoin Unlimited proposed letting miners negotiate the ceiling among themselves.
The other side had a different plan: fix transaction malleability, get a capacity increase as a side effect, and do it without a fork that could strand anyone. Segregated Witness shipped in Bitcoin Core 0.13.1 in October 2016 and began signalling on November 15, 2016 under a rule requiring 95 percent of mined blocks to agree within a year. It stalled around 30 percent through the winter, because miners had signed a deal in Hong Kong in February 2016 that tied their support to a hard fork nobody had delivered.
Two moves broke the deadlock in 2017. The New York Agreement of May 23 gathered 58 companies representing roughly 83 percent of hash power behind SegWit plus a 2 MB hard fork three months later. Simultaneously, a user activated soft fork, BIP148, set August 1 as the date on which participating nodes would start rejecting any block that did not signal SegWit. Faced with a possible chain split driven by users rather than miners, the industry adopted a bridging proposal, BIP91, which locked in on July 21, 2017 and forced signalling with days to spare. SegWit locked in during August and took effect at the end of that month. The hard fork half of the New York Agreement was abandoned on November 8, 2017 for lack of support.
Why this matters when you buy bitcoin
The block size war produced three things you will run into on your first purchase, whether or not anyone explains them.
The first is a family of similarly named assets. Bitcoin Cash forked away on August 1, 2017 with a larger limit and then split again on November 15, 2018 into Bitcoin Cash and Bitcoin SV. Exchanges list all of them, tickers sit next to each other in the same dropdown, and web domains carrying the word bitcoin have promoted the forks as the genuine article. Buy the ticker BTC, check the deposit network before sending anything, and treat any site arguing about which chain is real as an advertisement.
The second is your fee bill. The winning side kept base layer capacity conservative, so a busy mempool raises the price of a transaction rather than expanding to fit. That is why the fee rate you pick when you withdraw from an exchange is a real decision, and why the Lightning Network exists at all.
The third is a governance fact worth more than either. Miners with 83 percent of hash power signed an agreement and did not get their hard fork, because exchanges, wallet developers, and node operators declined to follow. If you run a node, you enforce the rules you chose; if you hold coins on someone else's platform, that platform speaks for you.
August 1, 2017, and the coins you held
The August 1, 2017 chain split is the clearest illustration of what custody means, and thousands of people learned it that week.
Block 478,558 is the last one both chains share, and after it the histories diverge. Every balance existing at that moment was now spendable on two chains, so anyone holding their own keys held both, whether or not they wanted the second one. Anyone holding coins on an exchange got whatever the exchange decided: some credited the forked coins promptly, some took months, and a few never did.
Claiming came with a trap. Moving forked coins usually meant importing a seed phrase into unfamiliar software, and a wave of tools appeared offering to do it for you. Some were straightforward theft. The safe pattern, then and now, is to move your bitcoin to a fresh wallet with a new seed before exposing the old seed to anything, and to accept that a few hundred dollars of forked coins is not worth risking the stack.
Block size war vs a hard fork
The block size war was a political argument; a hard fork is one of the tools it was fought with. Hard forks are routine in software and unremarkable in most blockchains, and the technical operation of loosening a rule so old nodes reject the result carries no drama by itself. What made this a war was the question underneath: whether a coordinated majority of miners and companies could change bitcoin's rules over the objections of the people running the software. The hard fork happened, on August 1, 2017, and the chain most people call bitcoin is the one that did not take it.