The cache's won (curPrice>=0.5 at pull) counts 50/50 refunds as wins for
BOTH sides — 521 of 2,128 chain-checked follow-set markets (24%) were
refunds, which is where the whales' 92-100% displayed win rates came from.
- live/payouts.py: resolutions table in cache.duckdb, filled from the CTF
contract's payout vectors (batched+paced JSON-RPC, ~12 calls/s free tier;
resolved rows immutable, unresolved recheck 6h, RPC failures never cached).
truth(cond, asset) -> 1/0/0.5/None; refunds need no asset side.
- validate_timing: every displayed stat (conv/conv30/all-time/realized/copy
replay) settles at truth; refunds count as neither W nor L, P&L is
size*(wp-p)/p; new conv_ref/conv30_ref/all_ref feed fields.
- trust.conviction_record: optional truthfn — the selection gates
(trust_wr/trust_roi) no longer select on refund inflation.
- portfolio.py: replay pays wp (refunds 0.5/share, was 1.0).
- conviction_scan: documented as the (refund-inflated) candidate layer;
final selection re-judges against truth downstream.
Validation: 0x4bFb-whale conv 174-16 91.6% $1.61M -> 34-16 +140ref 68%
$214k, and truth-adjusted all-time P&L now sits within ~16% of lb-api's
PM P&L (was 7x apart); LSB1 (0 refunds) byte-identical, its P&L matches
PM P&L to 0.07%.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Two data bugs were hiding the best copy targets (FINDINGS 'The holder
blind spot'):
* live/trust.py (new): only score rows whose res_t matches the market's
consensus resolution time across >=2 wallets, pulled after resolution,
resolved != False. Kills the res_t=ts fallback poison that let scalpers
masquerade as 99%-win holders (ArbTraderRookie's rows were 100% this).
* conviction_scan.py: trusted rows only; p80 conviction cutoff from the
train window only (look-ahead leak); new gates z_all>2 (whole-book z,
~doubles pooled forward copy-ROI) and median conviction stake >=$50
(dust filter). 55 wallets selected, forward 30/38 profitable, +21.4%
pooled (was 284 selected, +16.0%).
* validate_timing.py: the held-edge gates now read the trailing-90d
trusted conviction record via trust.conviction_record instead of the
replay's held leg, which is structurally ~all-unresolved for week-lead
holders (whale 0x73afc816 showed 'held 0-0, 21 unresolved' and was
rejected at 100% fwd win). 25 copy-positive holders now, including
Stavenson (51-0), the whale (112-4) and iohihoo (98-10).
* cache.py: query() helper so trust.py shares the single in-process
connection instead of fighting the single-writer lock.
* README: gotcha 8 (fake res_t/won) + candidate next data sources
(Goldsky pipelines, PolymarketData order books, Pinnacle CLV,
Polysights); FINDINGS: dated section correcting the iohihoo/ArbTrader
scalper-trap verdicts as winner=False-bug artifacts.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- copybot: model Polymarket taker fees (V2, since 2026-03-30: shares*rate*p*(1-p),
sports 0.03) on every paper/live fill; track fees_paid in state + feed; settle
P&L nets the entry fee. publish_feed now commits state + fills ledger with the
feed (autostash rebase) so the local poller is the sole state writer.
- validate_timing: copy_pnl/held_pnl are fee-aware -> sharp selection now requires
clearing real copy costs; conv stats + lead profile use resolved bets only.
- conviction_scan/skill: exclude res_t>now rows (early-sold positions in
unresolved markets scored at curPrice - a mark, not an outcome; was ~5% of the
June test window at a 72% pseudo-win rate).
- portfolio: skip unresolved rows (stake used to vanish from equity); missed bets
of kind=open no longer KeyError - mark-to-market hypothetical P&L.
- collect: cap stale refreshes at STALE_CAP=2500/run (bulk-aged pool turned the
daily refresh into a ~40h pull holding the DuckDB lock).
- Actions cron disabled: GitHub ran */5 every ~1.5-2.5h and the 10-min stale
window skipped the rest -> 1 of ~104 qualifying buys copied. launchd --poll 60
is now the runner; workflow stays as manual backstop.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Swap the flat $200 conviction cutoff for a per-wallet percentile (top
20% of each wallet's own stake sizes) everywhere it was used:
- cache.py: canonical CONV_PCTILE=0.80 + conv_cutoff() helper (matches
the dashboard's pctl: filter >0, sort, linear interp)
- conviction_scan.py: per-wallet quantile_cont(size,0.8) in SQL, was
`size >= 200`
- validate_timing.py, pnl_focused.py: use cache.conv_cutoff
Rationale + validation: p80 reproduces flat-$200's win-rate lift on the
sharps while adapting to scale (a whale's $200 isn't conviction, a
minnow's is). Re-running the pipeline under p80: scan finds 218 profile
wallets (was 69), forward 62/83 profitable (p~0), +16% pooled ROI — edge
persists out-of-sample. Regenerated conviction_wallets.json /
watch_sharps.json; docs updated. skill.py/strategy.py/insider.py
untouched (score over all bets / size as copyability heuristic only).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
conviction_scan.py finds wallets whose high-conviction (>=$200) bets win on
uncertain (~0.4-0.6) markets — real edge, not favorite-riding — trained pre-June
and validated June (25/37 stayed profitable forward, p=0.024). validate_timing.py
applies the entry->resolution lead-time gate that separates copyable sharps from
uncopyable insiders: of 69 matches, 21 were insiders (lead <6h), leaving 23
validated copyable sharps (watch_sharps.json) now shown on jaxperro.com/trading.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
jaxperro