How many dollars one transparent Zcash transaction moves — the mean, the median, the single most common amount, and the long upper tail (p90, p95, p99, p99.9), day by day. And the shape underneath: how many transactions are small, and where the money actually is.
Every transaction in the range, sorted into size bands a quarter of a decade wide. The top panel is each band's share of transactions; the bottom is its share of dollars moved. Two panels, one x-axis: where the two disagree is the whole story of a skewed distribution.
The line is the median transaction that day, in dollars at that day's close; the shaded bands are the middle 50% (p25–p75) and the middle 80% (p10–p90). The dashed line is the mean. Log scale: sizes span from cents to millions, and a linear axis would draw every ordinary day flat on the floor.
The size that only 10%, 5%, 1% and 0.1% of the day's transactions exceeded, with the median for scale. This is where most of the value moves: a handful of transactions at the top of the day routinely outweigh all the rest put together.
Transactions per day, and the dollars they paid to transparent outputs. Two panels rather than two y-axes: a count and a dollar total are different quantities.
One data point is one transaction: the total it paid to transparent outputs, priced at that day's ZEC close. Coinbase (mining reward) transactions are left out — they are issuance, not anyone moving money.
Change is included, and that makes every figure here an upper bound on what was paid. Spending a 100 ZEC coin to pay someone 1 ZEC sends 99 ZEC back to the sender as change, and the chain does not mark which output is which. So that transaction reads as ~100 ZEC moved. Every method of guessing the change output is a guess, and a guess baked into a stored number cannot be checked afterwards — so this page reports what is provable: how much value moved through transparent outputs, transaction by transaction.
Three views, chosen with the chips above. Transparent → transparent is transactions with transparent inputs and nothing going into or out of a shielded pool: coins moving between transparent holders. Leaving a shielded pool is transactions that release value from a shielded pool to transparent outputs. Everything also includes shielding transactions, whose only transparent output is usually their change — which is why it is not the default.
The mode is an exact amount, and that is what makes it meaningful. A dollar figure is continuous and has no useful mode, but a ZEC amount is a whole number of zatoshi, and automated payers repeat one amount exactly. When the mode covers a large share of a day, that day's median describes a robot, not a crowd.
Range figures are labelled for what they are. A percentile of several days pooled together cannot be rebuilt from per-day summaries, so the cards report the median across days of each daily figure — the typical day. The mean, counts and totals are exact over the whole range. The shape chart groups each day's transactions into size bands before pricing them, so a transaction can land up to half a band (about 1.3×) from its exact dollar size; it is a shape, and no percentile is read off it.
Shielded transactions hide their amounts, so this page cannot see them at all. For how much ZEC sits shielded, see Shielded Pool Balances. Source: every block, via a full node.