The privacy page made the case that financial privacy is worth having, and that on Zcash it
is a property you keep by how you use the tool — not a switch that is simply on. This page
is the honest other half of that story. Privacy technology is real, but it is not automatic,
and it has not always lived up to what was promised for it.
It helps to keep three different things apart, because they carry very different lessons.
A privacy tool can be promised but not delivered — marketed as private without really
being so. It can be made, then defeated — genuinely private, until someone finds a way
to undo it, usually by analysing how people use it or by watching the network underneath,
not by breaking the mathematics. And it can be interrupted while a bug is fixed — working
as intended, until a separate flaw forces the team to switch the private system off until a
repair ships. Most of what follows is the third kind: bugs that researchers found and teams
fixed, which is how security is supposed to work. The point of cataloguing it is the opposite
of fear — a reader who knows how privacy has actually come under pressure is harder to mislead
about what any coin can promise.
1. Promised, but not delivered
The clearest version of "it didn't do what the label said" is a tool whose privacy was thinner
than its marketing. Verge (XVG) is the stock example. It marketed itself on the word
Privacy, built around its Wraith Protocol (launched at the
end of 2017), which routed a user's connection through the Tor and I2P networks to hide their
IP address. The gap was in what that covered. Verge runs a
transparent ledger: its transactions are
publicly viewable on-chain, as in Bitcoin. It offered an optional "stealth" mode said to hide
transaction amounts, but its privacy was a set of opt-in layers bolted onto that transparent base
— not the mandatory, on-by-default on-chain confidentiality a shielded coin provides.
Marketed as anonymity, what shipped was narrower than the word suggested — a reminder to check
what a "privacy coin" actually hides before trusting it with anything.
2. Made, then defeated
Here the privacy was genuine and the cryptography held — but it could still be undone from the
outside. These are the cases the privacy page's transaction hygiene section was built on:
real privacy is lost through careless use or through the network layer, not through broken
maths.
Zcash — linked by usage. Zcash's shielded cryptography has never been broken, yet privacy
can leak through how the tool is used. In
On the linkability of Zcash transactions (Quesnelle,
2017), the author linked roughly 31.5% of all coins sent into the shielded pool by spotting
"round-trips" — the same, or nearly the same, amount moved into a shielded address and back
out to a transparent one a short time later. The next year,
An Empirical Analysis of Anonymity in Zcash
(USENIX Security, 2018) generalised the point: simple heuristics based on identifiable usage
patterns can sharply shrink the anonymity set — the crowd a payment hides in. Neither broke
the maths; both defeated careless use, which is exactly why the hygiene on the previous page
matters.
Monero — watched at the network layer. Monero hides every transaction's sender, receiver,
and amount on-chain, for all users by default, and that cryptography has held up too. The
pressure came from two other directions. An early study,
An Empirical Analysis of Traceability in the Monero Blockchain
(Möser et al., 2018), showed that weak "decoy" selection let an analyst guess the real coin
in a transaction with around 80% accuracy — a weakness Monero
largely closed
by changing how decoys are chosen and raising the mandatory ring size. The live pressure is
now at the network layer, not the cryptography: an entity dubbed
LinkingLion, active since 2018, has been
probing nodes to tie transactions to IP addresses, and a widely-reported 2024 leak appeared
to show the analysis firm Chainalysis operating Monero nodes for the same purpose. Monero's
developers track this as the public remote-node problem:
it records who connected, not what they sent, and the countermeasure is the network-layer
hygiene the privacy page closed on — route through Tor, or run your own node.
Grin and Dash — read from the transaction graph. The same idea — reconstructing who paid
whom by watching the network — has been demonstrated against other designs, with the projects'
teams contesting how much it proves. A researcher reported linking
96% of Grin transactions
(Bogatyy, 2019) by running listening nodes on Grin's Mimblewimble network; Grin's developers
replied
that this was a long-documented limitation of Mimblewimble, not a break of it. And
CoinJoin in the Wild
(Deuber & Schröder, 2021) found that over 40% of Dash's optional PrivateSend mixes could be
de-anonymised. The recurring lesson is the privacy page's: privacy that is optional, or that
leaves a usable pattern, is privacy you can lose.
3. Interrupted while a bug was fixed
The last kind is the most common — and, read carefully, the most reassuring. Here privacy went
offline not because it failed, but because the responsible response to a different bug was to
switch it off until a fix shipped.
Zcoin/Firo and PIVX — Zerocoin switched off. In 2019 a
flaw in the Zerocoin protocol let an
attacker forge the zero-knowledge proofs Zerocoin relied on and mint counterfeit coins. Zcoin
(now Firo) reacted by turning Zerocoin off in April 2019 and migrating users to its newer Sigma
(and later Lelantus) protocol — switching the feature off stopped the counterfeiting. PIVX had
already disabled its zPIV in March 2019,
before this libzerocoin flaw was discovered, so it was never exposed. Either way, the privacy
those coins offered was suspended until a replacement shipped.
Zcash itself — Orchard paused. Zcash has been through the same kind of episode, and
including it matters: the honest version of this page cannot make an exception for the home
team. The full history is on the timeline page; what counts here is what those incidents meant
for privacy. A
2018 counterfeiting flaw
in the original Sprout pool was fixed quietly before it was disclosed; in
March 2026 a
node-software bug touching the deprecated Sprout pool was patched within days. The most serious
surfaced at the end of May 2026, when researchers found a
soundness bug in the Orchard pool
that could have allowed undetectable counterfeit ZEC; the June response was an
emergency soft fork
that disabled Orchard for about a day, until the NU6.2 upgrade shipped a corrected circuit —
taking the shielded pool, and so its privacy, offline to protect the integrity of the money.
Throughout, Zcash's turnstile rule — a pool can never pay out more than was paid into it —
capped how far any such bug could distort the supply.
What the pattern teaches
Put the three side by side and a single lesson falls out, and it is not "privacy doesn't work."
The promised-but-not-delivered case says: check what a tool actually hides, not what it is
marketed as. The made-then-defeated cases say: real privacy can still be lost through careless
use or through the network underneath it — which is why the privacy page spent its time on
transaction hygiene and on the separate, network layer that Tor and mixnets address. And the
interrupted-while-fixed cases say something almost steadying: when a serious bug turns up in a
privacy system, the responsible move is to disclose it, switch off what is unsafe, and ship a
fix — even at the cost of taking privacy temporarily offline. Privacy is not a switch that is
permanently on. It is a property a sound design protects, an honest team defends, and a careful
user keeps.
Firo, Zcoin: 2019 in Review and PIVX,
FAQ on Zerocoin and PIVX — the 2019 Zerocoin
cryptographic flaw, Zcoin's disabling of Zerocoin in response, and PIVX's zPIV already
disabled before the flaw was found.