For the frozen ten-hour window, this on-chain identity did little besides maintain DeepBook order-book activity. It recorded 109,029 cancellations and 74,862 limit-order placements among 184,205 transactions with recognized transaction-level meaning. The near-zero median gap and 466 ms p90 gap are consistent with automated order maintenance, but the evidence does not establish who operated the address or whether the activity was profitable.
A very fast rhythm
These figures use the complete history from 2026-07-19 13:01:48.968 UTC through 23:37:42.097 UTC. The address submitted 196,441 stored transactions, with a mean gap of 194 ms and a median gap of 0 ms. Its busiest minute was 18:08 UTC, with 1,015 transactions; the longest idle period was only 39.6 seconds. The longest uninterrupted run of successful transactions was 2,136.
The repertoire barely changed by hour. Cancellations remained the leading activity in every hourly bucket, ranging from 52.7% to 58.5% of each bucket's stored transactions. Placement activity was consistently second. That stability, combined with the cadence, makes this look like a continuously running order-maintenance process rather than occasional manual trading.
What the calls actually show
All 109,029 cancellations used the same DeepBook owner-proof pattern. Placement activity was slightly more varied: 52,858 ordinary owner-proof placements and 21,977 transactions that reused one prepared proof across bounded runs. Only 51 transactions used a separate owner-authorized placement attempt pattern, while 263 transactions had recognized calls without an asserted behavior.
In AbxeSen8, the transaction detail describes a DEEP/SUI order cancellation at a 0.02311 SUI-per-DEEP resting price, with 510,410 DEEP recorded as resting size removed from the book. It explicitly distinguishes that resting size from an execution result. In DHYQcvzC, one proof was reused across three placement calls: 8,920 NS at 0.01686 SUI per NS, 9,582 NS at 0.01692, and 43,059.6 NS at 0.0171. The detail describes the supported outcomes as leaving those orders on the book, but these records do not establish later fills or economic benefit.
The underlying shape is concrete: generate an owner proof, then call DeepBook's pool::cancel_order; the three-order variant generates one proof and calls pool::place_limit_order three times. Every transaction called the DeepBook revision 0xcaf6ba059d539a97646d47f0b9ddf843e138d215e2a12ca1f4585d386f7aec3a.
Failures and cost
Across the complete history, 14,364 transactions failed. Of the 184,205 transactions with recognized meaning, 2,128 failed; another 12,236 failed before a recognized call could be described. The largest named stopping point was big_vector::leaf_remove, with 1,974 failures. Smaller groups stopped at balance_manager::withdraw_with_proof 86 times and order_info::assert_execution 68 times.
The sender itself paid every transaction's gas; none were sponsored. Gas evidence records 175.897744936 SUI out and 18.167980612 SUI in, for a net gas movement of 157.729764324 SUI out. Separate holding evidence records 159.732393220 SUI out across 719 transactions. Those figures belong to different evidence categories and must not be added together.
The final burst
The detailed stream is capped at the latest 500 transactions, so it is a different scope from the complete-history aggregates. Those 500 records span only 51.8 seconds, from 44itS71w at 23:36:50.271 UTC to AbxeSen8 at 23:37:42.097 UTC. They contain 291 cancellations and 209 placements; all 500 have an authored purpose explanation and none failed. The stream's reported net gas movement was 0.434838700 SUI out.
This address is best understood as a highly regular DeepBook order-maintenance identity: fast, repetitive, and narrowly focused. The evidence supports that character within this frozen window, but it does not identify an operator, prove a bot, establish fills, or show profit.
Further reading: AbxeSen8 for a concrete cancellation, DHYQcvzC for the repeated three-order placement, 44itS71w for the first record in the capped latest-500 stream, and DeepBook revision 0xcaf6ba059d539a97646d47f0b9ddf843e138d215e2a12ca1f4585d386f7aec3a for the mechanism.