A shield-bound transparent hop is one transparent coin that arrived, sat for a while, and was then spent by a transaction that shielded it. Because a coin is spent exactly once, that is the whole story of that coin: it landed in the open, and the next thing that happened to it was going private. This page asks how big those hops are — not how many ZEC moved in total, but how much moved at a time.
The line is the median hop that day; the shaded bands are the middle 50% (p25–p75) and the middle 80% (p10–p90) of that day's hops. Dots mark the highest and lowest medians in the range and the latest one. Log scale, because hop sizes span five orders of magnitude and a linear axis draws every ordinary day as a flat line on the floor beside one spike. Hover any day for its numbers.
Two panels rather than two y-axes: a count and a volume are different quantities, and pinning them to one plot on separate scales invents crossings that are an artefact of where each axis was put.
The share of each day's hops that were all exactly the same amount. A day near the top is not a sample of many people's decisions — it is one automated process repeating itself, and its median describes that process and nothing else. Hover a bar for the amount.
Chain shape only — no address data anywhere in this page. A hop is identified by what happened to a coin (it was spent into a transaction that shielded value), not by who held it. Everything below follows from that one limitation.
The reading most people want is “an exchange withdrawal or a deposit being made
private”, and that is a reasonable guess — but it is a guess, and this page will not make it for you.
A coin arriving transparent and leaving shielded is equally consistent with a withdrawal being secured, a holder
tidying up their own wallet, a payment received and put away, or a service moving its own float. We have no
counterparty, so the schema, the API and every label here say what is provable — transparent in, shielded
out — and stop there.
Direction is not observable either. Nothing in a hop says whether the person behind it is buying or selling. The tiles say moved, never bought or sold, and that is not coyness: the information is genuinely not on the chain.
Two things you can switch off, and why they are switches rather than arithmetic. Sweeps are hops whose spending transaction was sweep-shaped (≥15 inputs, ≤2 outputs) — the signature of something automated gathering many coins at once. The repeated amount is the single exact size that occurred most often that day; dropping it removes a payout robot that a sweep filter can miss entirely, because a robot paying one coin in and one coin out is not sweep-shaped at all. Both matter: on 2026-09-10, 8,000 of that day's 10,614 hops were one weekly process paying exactly 0.125 ZEC, which on its own moves the day's median from 1.25 ZEC to 0.125 ZEC. Neither can be subtracted from a number already computed — a median of what is left is not a function of the median of everything — so each combination is computed separately, per day, with the exclusion already applied. Whatever a switch removes is still reported beside the chart, because a number taken out of a distribution has to stay visible.
The unit is a coin, not a person and not a transaction. One shielding transaction can spend many transparent outputs, and the source table stores no transaction id, so a hop here is one output. The sweep switch is the available proxy for “was this one coin or a batch”, not a true per-transaction view.
Multi-day figures are labelled for what they are. A percentile of several days pooled together is not recoverable from per-day summaries, so the range summary reports the median of the daily medians — the typical day's typical hop — under that name. Counts and totals are plain sums and are exact.
Built from the same classified-spend data behind Transparent Activity, which answers the volume question this page answers the size question for. Source: every block, via a full node.