Skip to content
Ascertain
Tokens
Replay on production data

Real-token results.

The engine replayed on 2 real HyperEVM mainnet tokens. This is a partial analysis: holder integrity only. Volume and market-cap signals need a swap adapter per DEX and are not computed yet. Results describe observable patterns, never intent.

2
Tokens replayed
8
Clusters flagged
0/8
Hand-checked
pending
False positives

Hand-check is in progress. A flag counts as a false positive only after someone opens the wallets on the explorer and rejects it; until then no false-positive rate is claimed.

TokenVerdictHoldersTop-1 / top-10ClustersInfrastructure excluded
HPL
0xbd6dab…cac91b
WAIT4,428
funding checked: top 1,000
17.3% / 59%310 holders, 12 funders
HWAVE
0x0e01e3…22e79d
WAIT3,056
funding checked: top 1,000
10.5% / 63%56 holders, 8 funders
Flags to hand-check

Every cluster, with links.

Method and limits

  • All ERC-20 transfers since contract creation were downloaded and balances rebuilt. Receipts re-fetch totalSupply, balances and nonces at the snapshot block, and funding transactions by hash.
  • Contracts among the top 200 holders (pools, vaults, bridges) are excluded as infrastructure. Funders that are contracts or very busy accounts (over 20,000 transactions) are ignored as exchange-like or router-like. Both rules are heuristics and are listed per token in evidence/real/.
  • Funding links use the first inbound native-token transfer of the top 1,000 holders. Funding through exchanges, mixers or contracts is invisible.
  • Receipts were verified through the same RPC provider as the indexer, so they are a second request, not a second provider.
  • No volume, market-cap, wash or same-block findings are published for real tokens yet.