From c20ac0e78f96df80b8f920f76fb0a7f900dd248b Mon Sep 17 00:00:00 2001 From: Theodore Song Date: Tue, 18 Aug 2026 12:50:42 -0400 Subject: [PATCH] Cluster adaptive evidence by market --- README.md | 32 ++++++++++++++--------- index.html | 77 ++++++++++++++++++++++++++++++++++++++---------------- sw.js | 2 +- 3 files changed, 74 insertions(+), 37 deletions(-) diff --git a/README.md b/README.md index f396fce..0fc3750 100644 --- a/README.md +++ b/README.md @@ -15,7 +15,7 @@ https://polymarket-site-eta.vercel.app/personal.html The site fetches live Polymarket markets, generates agent suggestions, lets you run frequent paper cycles, and syncs the shared arena state through Neon or -Vercel Blob. Build 50 also installs an offline app shell and caches timestamped +Vercel Blob. Build 51 also installs an offline app shell and caches timestamped market snapshots. During an outage, cycles continue locally; cached entries are allowed for 90 minutes, older snapshots become mark-only, and all cached data expires after 24 hours. @@ -27,7 +27,7 @@ The learner shrinks small samples toward neutral, caps sizing changes to 15% of candidates for deterministic exploration so a stale regime cannot become permanent. -Strategy 46 treats each binary stake as capable of falling to zero even when the +Strategy 47 treats each binary stake as capable of falling to zero even when the 18% stop cannot fill. New core positions are capped at 2.5%-4% of equity and aggressive positions at 3%-5%, with lower limits for near-term, extreme-price, reversal, and fast-moving setups. Oversized positions inherited from older @@ -65,7 +65,7 @@ negative in all three chronological segments. Reversals averaged -3.69%, with a market-clustered 90% interval entirely below zero. Crypto and Sports were also negative but covered only three and five markets. The 24-hour cohort improved to -0.82% row mean and +1.31% market mean, with no rule robustly negative across all -segments. Strategy 46 therefore keeps reversal entries observation-only until +segments. Strategy 47 therefore keeps reversal entries observation-only until their recent signal and quality cohorts independently earn promotion, retains their signals for paper grading, and evaluates adaptation at 24 hours. @@ -74,7 +74,7 @@ markets with no failures and produced 3,597 net-of-cost 24-hour outcomes across 142 markets. No tested follow or fade rule was robustly positive. Crypto trends averaged -3.83% per observation and -3.99% per market; Sports trends averaged -5.44% and -6.61%. Both stayed negative in every chronological segment and their -market-clustered 90% intervals were entirely below zero. Strategy 46 therefore +market-clustered 90% intervals were entirely below zero. Strategy 47 therefore keeps Crypto and Sports trends observation-only while continuing to grade them. A corrected 200-market audit paged through 197 markets with usable history and @@ -83,23 +83,23 @@ segment and averaged -4.13%. Sports trends were negative in train and test and averaged -3.53% at 72 hours. Politics trends were the sole cohort with positive row-level returns in all three 72-hour segments, but its market-cluster interval still crossed zero; that supports a longer hold test, not a larger entry bet. -Strategy 46 gives Politics trend positions that 72-hour observation window before +Strategy 47 gives Politics trend positions that 72-hour observation window before ordinary signal exits. Stops, profit locks, settlement handling, and risk-budget reductions remain immediate. -Strategy 46 also subtracts a half-cent round-trip cost when grading each live +Strategy 47 also subtracts a half-cent round-trip cost when grading each live walk-forward signal. Confidence uses the largest independent matching bucket, not the sum of five overlapping feature buckets, and evidence from older engine versions is down-weighted. This prevents a handful of duplicated observations from authorizing larger positions or hiding a modest negative regime. -Strategy 46 adds uncertainty-aware promotion and demotion. A matching setup must +Strategy 47 adds uncertainty-aware promotion and demotion. A matching setup must accumulate at least eight effective observations and agree across at least two feature views before repeatable positive evidence can increase size or repeatable negative evidence can block a new entry. Mixed evidence stays close to neutral instead of being mistaken for an edge. -Build 50 enforces the documented offline boundary end to end. Cached snapshots +Build 51 enforces the documented offline boundary end to end. Cached snapshots under 90 minutes old may continue paper execution. Older snapshots remain usable for valuation and chart snapshots for up to 24 hours, but cannot trigger entries, stop-losses, gain-stops, risk rebalances, settlements, or policy exits. Network @@ -112,12 +112,12 @@ adaptive baselines, pending signal grades, and trade evidence remain in one stra lineage until the actual entry, sizing, or exit logic changes. Legacy build 40 and 41 records are migrated into the same strategy lineage without losing evidence. -Build 50 independently refreshes markets for matured pending signals that have +Build 51 independently refreshes markets for matured pending signals that have left the current top-500 activity scan. Unavailable markets remain queued for a bounded retry window. This prevents activity-rank survivorship from deciding which wins and losses reach the adaptive calibration ledger. -Strategy 46 coordinates high-risk exploration globally. Near-term, extreme-price, +Strategy 47 coordinates high-risk exploration globally. Near-term, extreme-price, and other gap-prone positions may be held materially by only one agent, while ordinary independently confirmed markets retain the two-agent cap. The robustly negative Sports- and Crypto-trend cohorts cannot enter through exploration. @@ -139,7 +139,7 @@ whole-event YES or NO bundles using executable best asks/bids, per-leg costs, an a minimum-liquidity requirement. The 500-event audit found 33 complete liquid negative-risk events and zero positive worst-case bundle returns after costs. Midpoint price sums sometimes looked attractive, but executable spreads removed -the apparent edge, so Strategy 46 does not pretend those snapshots are arbitrage. +the apparent edge, so Strategy 47 does not pretend those snapshots are arbitrage. The expanded event-clustered run loaded history for 498 of the 500 highest-volume resolved markets with no fetch failures. No side, price band, category, trend, @@ -150,18 +150,24 @@ static settlement-direction boost from this audit. The earlier 200-resolved-market audit found short-dated NO entries strongly negative, but the 500-market rerun did not reproduce that loss in its newer test -segment. Strategy 46 therefore treats the result as a provisional prior instead +segment. Strategy 47 therefore treats the result as a provisional prior instead of a permanent ban: NO entries with 21 days or less remain observation-only until the recent walk-forward calibration promotes their matching side and duration cohorts. Exact numeric-range contracts are excluded from new entries because a settlement jump can pass directly through an 18% stop; the live audit found that this failure mode caused the largest latest-day loss. -Strategy 46 also excludes path-dependent barriers such as "reach $66,000," "hit +Strategy 47 also excludes path-dependent barriers such as "reach $66,000," "hit $90," and "dip to $62,000." These contracts can resolve abruptly as soon as the barrier is touched, so a later hourly stop cannot reliably cap the loss. Fixed-date level questions such as "above $66,000 on August 23" remain eligible. +Strategy 47 clusters live walk-forward observations by market before calculating +confidence. Multiple six-hour snapshots of the same contract are averaged into +one effective market outcome, so one repeated winner or loser cannot promote or +demote an entire feature cohort. Promotion still requires at least eight weighted +distinct markets and agreement across two feature views. + Paper accounts created with a password are also saved through the backend, so a user can log in from another device and see the same paper portfolio, activity, and value history. Passwordless paper accounts remain local-only. diff --git a/index.html b/index.html index 68d7efe..a32f112 100644 --- a/index.html +++ b/index.html @@ -341,7 +341,7 @@ footer{margin-top:34px;padding-top:22px;border-top:1px solid var(--border);color