What the Forks Tell Us About Changing Bitcoin

Between 2017 and 2018, several groups concluded that Bitcoin had taken a wrong turn and launched their own versions of it. Each kept the ledger history, kept the name, changed a rule they thought was mistaken, and argued that theirs was the real Bitcoin. It was the most direct experiment the asset class has ever run, and the results are unambiguous.

What happened

Measured against Bitcoin itself, from our forks page: Bitcoin Cash has lost roughly 96%, Bitcoin SV roughly 98%, and Bitcoin Gold effectively 100%.

The dollar figures are no gentler. Bitcoin Cash peaked near $3,900 in December 2017 and Bitcoin Gold near $450 in the same month; Bitcoin SV peaked around $440 a year later. All three are down more than 94% from those peaks, and Bitcoin Gold by more than 99.9%.

The listing record is its own verdict. Of the major venues in our data store, Bitcoin Cash still trades on two, Bitcoin SV on one, and Bitcoin Gold on none. The fork that fell furthest is also the one no exchange we track still carries — and its reported volatility figure, several hundred percent annualised, is measuring the absence of a market rather than the behaviour of one.

What was actually being disputed

The headline argument was about block size. Bitcoin's one-megabyte limit caps transactions per block, which caps throughput and pushes fees up when demand is high. Raising the limit is a one-line change and obviously increases capacity. The counter-argument was that bigger blocks make running a full node more expensive, and that a network only a handful of well-resourced parties can validate has given away the property that made it worth having.

Underneath the technical dispute was a governance question, and that is the one the market appears to have answered. If a rule can be changed by a determined faction, then no rule is fixed — including the one about twenty-one million coins. The supply cap and the block-size limit are enforced by the same mechanism: nodes rejecting blocks that break the rules. A network that changes a contested rule under pressure has demonstrated that it can.

The uncomfortable reading

The forks did not fail because their technical arguments were wrong. Larger blocks do increase throughput. Cheaper transactions are genuinely better for payments. On the merits of the specific change, the fork advocates had a real case.

They failed because they demonstrated that the thing they were forking could be forked. Once you have shown the rules are negotiable, you are asking people to hold an asset whose scarcity depends on continued agreement rather than on refusal to agree. That turns out to be the entire product.

It is worth being honest about the alternative explanation: network effects, liquidity and brand may simply have carried the day, and the same outcome might have occurred if the dispute had been about something trivial. Both stories fit the data, and a sample of three forks cannot separate them.

What the forks did to their own supply

One detail is worth stating because it undercuts a common defence. All three forks kept the twenty-one million cap. They were not experiments in easier money — they were arguments about throughput that left the supply schedule alone.

So the market's verdict cannot be read as a preference for scarcity over inflation, because scarcity was not what differed. What differed was whether the chain had demonstrated that its rules could be changed by a faction under pressure. On the only variable that was actually on the table, the market punished the change.

It is also worth remembering what the dispute produced on the winning side. Bitcoin did not simply refuse to scale — SegWit activated in August 2017, and the Lightning Network was built on top of it. The block-size limit held while throughput was addressed in a layer that did not require every node to carry more data. Whether that was the better engineering answer is arguable; that it was an answer is not.

Why the numbers are worth keeping

The forks are the closest thing Bitcoin has to a control group. They shared its entire history up to a specific block, then diverged on one variable. Nothing else in this asset class offers that.

They also stand as a check on a common argument. "Bitcoin is just software, anyone can copy it" is true and has been tested at the most favourable possible terms — copies that inherited the ledger, the holders, the exchange listings and the name. All three lost more than 96% against the original.

The related measure is Bitcoin dominance, which asks the same question across the whole asset class rather than just the forks. It bottomed at 32.1% on 2018-01-13, at the height of the period when the alternatives looked most credible, and had recovered to 57.1% by the October 2025 cycle peak — the first cycle top at which Bitcoin's share was higher than at the previous one.

Last updated 2026-08-10.