Calibration
Computed . Refreshed about every 5 minutes.
Read this first
Survival from Watch entry
Watch entries in the last 7 days: of those read at or after the horizon, how many read at or above the entry cap. A rate appears only at 30 or more readings.
Swipe for 6h and 24h →
Outcomes so far
Every rule version. An outcome with no winner verdict is unknown, never a non-winner. The 2h and 24h multiples and 10x so far are interim, not the winner label.
1552 outcomes · 311 resolved · 80 unknown · winners 24 / 231 · 10% (7.1–15%) · capturable 16 / 23 (1 not measured) · 30 resolved reached · time to 10x median 1h 00m · p95 2d 19h · n 171
Swipe for winners and the interim counts →
Lane transitions
Every lane change read on the tape, from the lane before it. Readings and distinct coins; counts only.
Last 24h (since )
Last 7d (since )
v4.6.0live
Scored on coins that became actionable while this version was live.
Thresholds written by ensureRuleVersion() on the first tick · live from · row written
5339 coins · 0 resolved · 0 never eligible · winners none yet · Sub-$100k kill: unmeasured (0 resolved)
Swipe for winners, false kill and forgone →
v4.6.0-shadow-watch-18hshadow
Scored on every coin it placed, for comparison. It never made the call.
v4.6.0 with the Graduated band closing at 18h (the v4.5.0 line): the undo shadow of D107. Only gradAgeH differs from live. · shadow, never live · row written
5339 coins · 238 resolved · 0 never eligible · winners 17 / 175 · 9.7% (6.2–15%) · 1 winners the Sub-$100k kill removed
Swipe for winners, false kill and forgone →
v4.5.0-shadow-wash-hourshadow
Scored on every coin it placed, for comparison. It never made the call.
v4.5.0 with the Wash volume numerator read on the last hour (over 1x cap) instead of the 24h pile (over 15x cap); pool and cap legs unchanged (F02-3). Only the wash check differs from live. Base moved to v4.6.0 (gradAgeH 12, D107). · shadow, never live · row written
6335 coins · 275 resolved · 0 never eligible · winners 18 / 206 · 8.7% (5.6–13%) · 1 winners the Sub-$100k kill removed
Swipe for winners, false kill and forgone →
v4.4.0-shadow-sized-watchshadow
Scored on every coin it placed, for comparison. It never made the call.
v4.4.0 with the v4.3.0 reading of Sized, too young (Watch, not Wait): the undo shadow of D86. No threshold differs from live. Base moved to v4.6.0 (gradAgeH 12, D107). · shadow, never live · row written
6594 coins · 284 resolved · 0 never eligible · winners 20 / 211 · 9.5% (6.2–14%) · 1 winners the Sub-$100k kill removed
Swipe for winners, false kill and forgone →
v4.1.0-shadow-clock-firstshadow
Scored on every coin it placed, for comparison. It never made the call.
v4.1.0 with the v4.0.0 gate order: clock before the dust, curve and off-quote stops (D68). Base moved to v4.6.0 (gradAgeH 12, D107). · shadow, never live · row written
6791 coins · 303 resolved · 0 never eligible · winners 24 / 227 · 11% (7.2–15%) · 1 winners the Sub-$100k kill removed
Swipe for winners, false kill and forgone →
v4.1.0-shadow-fdv-as-capshadow
Scored on every coin it placed, for comparison. It never made the call.
v4.1.0 with FDV read as cap by the size gates (v4.0.0 rule) (D71). Base moved to v4.6.0 (gradAgeH 12, D107). · shadow, never live · row written
6791 coins · 303 resolved · 0 never eligible · winners 24 / 227 · 11% (7.2–15%) · 1 winners the Sub-$100k kill removed
Swipe for winners, false kill and forgone →
v4.1.0-shadow-lottery-killshadow
Scored on every coin it placed, for comparison. It never made the call.
v4.1.0 with the Sub-$100k kill (v4.0.0 rule). Promote only at 30 resolved forgone-winner and false-kill outcomes (D70). Base moved to v4.6.0 (gradAgeH 12, D107). · shadow, never live · row written
6791 coins · 303 resolved · 0 never eligible · winners 24 / 227 · 11% (7.2–15%) · 1 winners the Sub-$100k kill removed
Swipe for winners, false kill and forgone →
Shadow rules beside live
Distinct coins placed per segment: the live version from stored ticks, each shadow version from its shadow placements. Counts only.
Swipe for every shadow version →
Coins v4.6.0-shadow-watch-18h would place differently: the latest live raw segment against the latest shadow segment. Nothing here moves a coin; the live rule set still decides.
Kills under v4.6.0-shadow-watch-18h: coins whose latest shadow placement is a kill, and of those resolved, the winners it would have removed. Promotion needs 30 resolved outcomes in the segment.
Swipe for false kill, forgone and status →
Coins v4.5.0-shadow-wash-hour would place differently: the latest live raw segment against the latest shadow segment. Nothing here moves a coin; the live rule set still decides.
Kills under v4.5.0-shadow-wash-hour: coins whose latest shadow placement is a kill, and of those resolved, the winners it would have removed. Promotion needs 30 resolved outcomes in the segment.
Swipe for false kill, forgone and status →
Coins v4.4.0-shadow-sized-watch would place differently: the latest live raw segment against the latest shadow segment. Nothing here moves a coin; the live rule set still decides.
No coin placed differently yet.
Kills under v4.4.0-shadow-sized-watch: coins whose latest shadow placement is a kill, and of those resolved, the winners it would have removed. Promotion needs 30 resolved outcomes in the segment.
Swipe for false kill, forgone and status →
Coins v4.1.0-shadow-clock-first would place differently: the latest live raw segment against the latest shadow segment. Nothing here moves a coin; the live rule set still decides.
Kills under v4.1.0-shadow-clock-first: coins whose latest shadow placement is a kill, and of those resolved, the winners it would have removed. Promotion needs 30 resolved outcomes in the segment.
Swipe for false kill, forgone and status →
Coins v4.1.0-shadow-fdv-as-cap would place differently: the latest live raw segment against the latest shadow segment. Nothing here moves a coin; the live rule set still decides.
Kills under v4.1.0-shadow-fdv-as-cap: coins whose latest shadow placement is a kill, and of those resolved, the winners it would have removed. Promotion needs 30 resolved outcomes in the segment.
Swipe for false kill, forgone and status →
Coins v4.1.0-shadow-lottery-kill would place differently: the latest live raw segment against the latest shadow segment. Nothing here moves a coin; the live rule set still decides.
No coin placed differently yet.
Kills under v4.1.0-shadow-lottery-kill: coins whose latest shadow placement is a kill, and of those resolved, the winners it would have removed. Promotion needs 30 resolved outcomes in the segment.
Swipe for false kill, forgone and status →
Baseline M0
Winner rate of all eligible coins: every lift is measured against it.
Overall 24 / 231 · 10% (7.1–15%)
Swipe for regime and winners →
Data health
Snapshot completeness
Scheduled observations captured, of those due (due time plus tolerance passed) for coins the store was tracking.
Swipe for captured and late →
Model ladder
Conditional counts on a chronological split, with an embargo between train and test.
train 161 until · embargo 7d 0h · test 0 from
Swipe for M0 and lift →