4 Oct source basis. The interactive Arc tape ends at 4 Oct 15:00 UTC (exclusive). It uses 334's isolated FINAL-free research-replica generation, delayed row-exact recheck and frozen-member continuity gate; the separate 4 Oct FINAL generation is excluded. The frozen 330c empirical hook-cost rule still excludes unmeasured hooked pools. A tied opposite-direction exhaustion event changed one unselected j30 theta=-1 overlay's cache order in 334/333; selected comparator overlays in 334 were exact, but that tie limits whole-roster prefix claims. The eight-wallet live comparison is frozen at 3 Oct 10:45 UTC; its figures are historical and are not a 4 Oct mark.

330ja · Arc recent-regime re-split

Exploratory recent-regime screen. Selection uses only mature entries from 26 Sep 12:00 to 28 Sep 00:00 UTC across 2,327,400 frozen books. The blue band begins untouched OOS on 28 Sep; cutoff is 4 Oct 15:00. The short screen and large search create high overfitting risk. The solo-window presets are labelled diagnostic.
Corrected Arc execution, frozen RH models. Primary results use empirical hooked-pool costs and the historical composer participation rule. A separate p95 unmeasured hook-cost sensitivity appears in the report.
Gray shading before 28 Sep is the selection seal; immature pre-28 entries are shown separately. Orange shading marks final wind-down: only complete OOS entry cohorts count in metrics, including their exits inside that shading. The fleet curve stays off the main chart.

Presets

Portfolio knobs

bankroll $ rejectclip
per-token cap $live exposure, all members
Arc participation
λ crowding00.51
3rd-party router fee / side00.5 pp1 ppsensitivity; our fleet pays 0
one entry / pool-minoffon · first member only
source windowfull frozen tape (continuous)original 18–25 source onlyoriginal forward source onlyall source segments
all scales ×applycopy link

Members

book (hover for id)scalesleeve cap $

Scale multiplies each frozen RH leg's Arc notional. Sleeve cap limits that member's live capital; blank = none. Solid lines = all tokens, dashed = hooked-only.

Entry view assigns eventual settled PnL to its entry time; diagnostic, not live equity.

Cumulative PnL by exit time — complete entry cohorts

Capital in use (entry notional of open positions)

How "Replace live fleet" works

Live trades are the real on-chain swaps the eight production Arc wallets made, read from the research replica and grouped into cells of (pool, UTC minute, side). Replacement estimates what would happen if this portfolio took their place.

On (replacement): in every cell, the first L dollars of our same-side flow (L = what the live fleet already bought, or sold, in that pool and minute) pay no new own impact — the price had already moved by that much in the history the backtest reads. Only impact above L is charged. Off (additive): our impact is charged in full on top of the live fleet's (the 005 convention) — that double counts the pool move the live fleet caused.

Limits: credit applies only to the same pool, minute and side. Opposite-side flow never offsets. Hooked-v4 pools in the corrected Arc books rarely overlap the live fleet, so the credit may be small. The fleet baseline is shown in the table below and is excluded from the main PnL chart.

Member contribution

Joint = admitted legs inside the portfolio after Arc participation, crowding, caps and live replacement. Selection and OOS columns use the new 28 Sep split.

Token concentration (top 12 by PnL)

Impact: paid vs offset by live flow

Daily PnL (by exit day, current settings)

Live fleet (research replica, read-only)

Method and receipts