Hashrate is not one number
A hash is not a fixed unit of work, so a hashrate league table across algorithms adds quantities that cost different amounts to produce.
Hashrate is reported the way temperature would be if Celsius and Fahrenheit shared a symbol. Two chains both quote a figure in hashes per second, the figures are compared, and the comparison is meaningless — because a hash is not a fixed quantity of work.
SHA-256 is deliberately cheap to compute, so a modern ASIC performs on the order of 1014 of them every second. Scrypt is memory-hard, and each hash costs far more silicon and far more energy. Putting the two totals side by side and calling the larger one more secure compares a count of cheap operations against a count of expensive ones.
So the column groups by algorithm. Within SHA-256 the five chains that use it are genuinely comparable; within Scrypt, the two are. Across the boundary the numbers are simply not the same unit, and the layout says so rather than inviting the arithmetic.
Borrowed work is somebody else's work
The second problem is whose hashes they are. A merge-mined chain is secured by miners who are paid by a different network and who submit the same work to both. The security is real — and it is contingent on economics the chain does not control, and it leaves the moment the parent chain's incentives change.
Rather than assume how much of a figure is borrowed, we counted. Dogecoin: 60 of 60 consecutive blocks carried an AuxPoW header, meaning none of the work was done for Dogecoin alone. Its own hashrate is nil, and the figure usually quoted beside it belongs to Scrypt parents — some of them other than Litecoin, since the total runs above Litecoin's own. Fractal Bitcoin is the mixed case: it does mine for itself, interleaving its own blocks with ones carrying Bitcoin's work, but publishes no split. Inferring one from a version bit whose meaning cannot be confirmed would be inventing a number, so its own share reads as unknown.
Chains in that position are shown with the borrowed figure, marked, and sorted last among the mined chains — never ranked on hashpower that is not theirs.
What would actually compare
Cost. Energy burned per unit time, or money spent per unit time, is a single unit that means the same thing on every algorithm and every chain. It is also much harder to source honestly, which is why nobody publishes it and everybody publishes hashes instead.
Compare within an algorithm, never across one. And rank a chain on the work done for it, not on every hash recorded against it.
Figures are from the data build of 21 August 2026 and move with every six-hourly refresh. The argument does not.