Extraction timing traces (L2→L1→exchange)
This evidence pack attempts to strengthen (or falsify) the on-chain leg of the yield-extraction thesis by linking a tighter, repeatable sequence than a one-off exit:
- Arbitrum
WithdrawStake(liquid LPT leaves BondingManager) - Arbitrum bridge-out (LPT burn via gateway router)
- Ethereum L1 escrow release (LPT transfer from the Livepeer L1 escrow)
- Tight-window routing into labeled exchange endpoints (best-effort; often via a repeatable second-hop EOA)
This is still not proof of delta-neutral hedging (the hedge is mostly off-chain). It is, however, a measurable “cashout routing + timing” fingerprint that is harder to explain as a single discretionary exit.
Summary
-
Generated:
2026-01-22T00:46:05.137369+00:00 -
Arbitrum RPC:
https://arb1.arbitrum.io/rpc -
Ethereum RPC:
https://rpc.flashbots.net -
L1 token:
0x58b6a8a3302369daec383334672404ee733ab239 -
L1 escrow:
0x6a23f4940bd5ba117da261f98aae51a8bffa210a -
L2 window:
5856381→423863641 -
L1 window:
14600000→24286807 -
Senders analyzed: 10
-
Burn (bridge-out) events: 85
-
Matched
WithdrawStake→burn (≤ 72.0h): 75 -
Matched burn→L1 escrow receipt (≤ 60.0d): 84
-
Matched L1 receipt→labeled exchange (≤ 72.0h forward, then ≤ 72.0h to exchange): 68
Sender table
Columns: Arbitrum bonded stake now (snapshot), number of bridge-outs, and how many cycles can be followed all the way to a labeled exchange endpoint with tight timing windows.
| Sender (L2) | Bonded now (LPT) | Burns | Matched withdraw→burn | Matched burn→L1 receipt | Matched receipt→exchange | Median burn→receipt (d) | Median receipt→exchange (h) |
|---|---|---|---|---|---|---|---|
0x3d6182c59dbbbbc648570762da316ac8404816ae | 0.000 | 16 | 16 | 16 | 11 | 8.08 | 0.19 |
0xc5519fd1129d6d22744e0ac491401fff45d26528 | 2,322,141.979 | 18 | 16 | 17 | 16 | 7.34 | 2.85 |
0x962b029508b1054e2af4184bbaeb5d0c796f7526 | 0.000 | 1 | 1 | 1 | 0 | ||
0x60b8eb0947d25194afa248b81a087214beec5cfe | 0.000 | 2 | 0 | 2 | 2 | 7.01 | 0.18 |
0xef83273cbd014c4ae7998467c422275a8b37827e | 0.000 | 1 | 1 | 1 | 0 | ||
0x543df23b9a881fbdbdec7ab3e90f3ff7b905068b | 0.000 | 1 | 1 | 1 | 0 | ||
0x86abf78ac7ef44423873dabff35fe3e462b1ff6e | 283,800.279 | 12 | 8 | 12 | 7 | 7.06 | 0.27 |
0x8d37a11a65b4f2c541af1312bc44f74a078160c7 | 0.000 | 1 | 1 | 1 | 1 | 10.68 | 0.39 |
0xadc6e5cbde4fbac7baf58a336edeab8590625baf | 0.000 | 2 | 2 | 2 | 0 | ||
0xa3bd517dcbdc063c4c24f0d9837bbc5ce869d092 | 592,264.609 | 31 | 29 | 31 | 31 | 7.04 | 0.19 |
Notes + limitations
- This report relies on the canonical Arbitrum bridge-out signature captured in
research/arbitrum-bridge-out-decode.json. - “Exchange” routing is label-set based (best-effort). The absence of a labeled exchange does not imply the tokens weren’t sold.
- Many flows route via one or more EOAs on L1 before an exchange deposit; this report only follows one intermediate hop.
Raw output: see research/extraction-timing-traces.json.