The Genesis Trail

Methodology

Data source

All queries run against a full cardano-db-sync PostgreSQL database: the canonical relational index of the Cardano chain, which exposes blocks, transactions, inputs (tx_in), and outputs (tx_out) as SQL tables. A hosted, fully-synced mainnet instance was used, so no local sync was required.

Why SQL and not a REST API: services like Blockfrost, Koios, or Kupo expose fixed endpoints (look up a known address / tx / asset). Discovering an unknown party by amount + cadence needs an arbitrary value-pattern scan across a whole year, which only direct SQL over db-sync can run.

Units & schema notes

  • Values are stored in lovelace: 1 ADA = 1,000,000 lovelace, so 20,000,000 ADA = 20000000000000.
  • A payment's recipient is an output address (tx_out.address).
  • A payment's sender(s) are resolved by joining each input back to the output it spends: tx_in → tx_out (tx_out_id, tx_out_index).
  • Wallet linkage uses the stake credential (stake_address.view) where present; Byron-era (DdzFFz… / Ae2…) and enterprise addresses have no stake key.
Transfers, not sales

The ledger records transfers: value moving from one address to another. It cannot distinguish a sale from any other move, because a sale (disposal for cash, or any change of beneficial ownership for value) settles off-chain. So this analysis cannot ascertain that any sale occurred. What it can do is locate transfers whose amount and timing correspond to the purported sales, and report that correspondence; reading those transfers as sales would need off-chain evidence we do not have.

The search ladder (most → least restrictive)

  1. Exact 20M single outputs in 2021: 41 found, but none formed a monthly series to/from any single party.
  2. Group by recipient / sender stake: no single stake recurred ~monthly (the candidates were stakeless Byron addresses, lumped together).
  3. Widen to a band and aggregate by month: for each address, sum receipts per calendar month and look for ~20M in many months. This surfaced a single address with ~20M in 6+ months.
  4. Drill the candidate: confirm the actual dates, amounts, and ~28-day gaps.
  5. Resolve senders: find the funding wallets.
  6. Recursive trace to genesis: see below.
A key lesson

The recurring receipts were 20,200,000 ADA, not exactly 20M, just outside a tight ±0.5% band. Aggregating monthly receipts (tolerant of amount drift and in-month splitting) is what revealed the series.

Tracing fund origin: the dominant-input walk

To test the "2.4B genesis holder" claim we walk the input graph backward with a recursive SQL query (WITH RECURSIVE). At each transaction we follow the single largest input back to the transaction that created it (the dominant flow of value) and ignore the smaller inputs, repeating until we reach a transaction with no inputs: a Byron genesis output.

2.46B2.4B180MIOG genesisno inputshop tx2.4Bhop tx180Mpayment tx20.2Mlargest input, followedsmaller inputs, ignoredwalk right → left: follow the dominant input each hop until a tx has no inputs (genesis)

Why this can be valid, and where it can mislead

It is a heuristic, not proof. Cardano's UTxO model commingles every input of a transaction into one pot before paying the outputs, so no input is earmarked to any particular output. Weigh both sides:

Why it's informative
  • It follows the bulk of the value: the path that actually carries the money.
  • It is deterministic: the same dominant-input walk always yields the same chain.
  • All nine payment chains, traced independently, converge on the same genesis output: convergence that strong is hard to explain by chance.
  • It is tractable where full multi-input taint analysis explodes combinatorially.
Why it isn't proof
  • A transaction commingles its inputs: "largest input" is a heuristic for which prior tx funded it, not a guarantee.
  • It ignores smaller inputs, which could be the true source of a given output.
  • It can be defeated by structuring: splitting funds so no input dominates, or padding a hop with a large unrelated input.
  • Change outputs and mixing can redirect the dominant path.
  • Deep chains can hit the recursion cap and stay inconclusive (see the consolidation address's depth-capped depositors).
  • Convergence corroborates origin of funds; it never establishes a real-world identity.

Read the verdicts on this site accordingly: we report what the dominant flow indicates, with the convergence of nine independent chains as the corroboration.