Task02f/control group and proxy refinement#6
Merged
Conversation
All trades in DB have outcome_index=1 (NO token). Raw VWAP was storing NO price. Fix: YES_price = 1 - VWAP. Result: Brier 0.5484 → 0.1347, calibration bins now monotone. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Pilot (event_resolved, n=725): median=-0.084, %pos=15.2% Control (unclassifiable, n=683): median=-0.043, %pos=21.4% Mann-Whitney p=0.000001, r=0.132, CI [-0.066,-0.023] entirely negative. Verdict: REVERSED SEPARATION — pilot ILS is significantly LOWER than null baseline. Positive ILS in event_resolved markets not elevated above unclassifiable baseline. Root cause: resolved_at-24h proxy structurally better for sports/ behavioral markets (resolution IS the event) than for political/ regulatory markets (news precedes resolution by days). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Key results on 725 pilot markets: 24h: %pos=15.2%, median=-0.084, %|ILS|>1=13.9% 6h: %pos=11.0%, median=-0.134, %|ILS|>1=19.3% 2h: %pos=0.0%, median=-0.332, %|ILS|>1=25.3% 1h: %pos=0.0%, median=-0.350, %|ILS|>1=27.1% Spearman ρ(24h,1h)=0.542 — moderate proxy sensitivity. Monotone increase 24h→1h: only 14/221 markets (6.3%). Epstein cluster highly unstable: AOC 24h=+0.933 vs 6h=-4.241. Interpretation: news released >24h before resolution; final 6h captures post-disclosure price noise, not pre-event leakage. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…let analysis Phase 3 of Task 02F: deep dive on the 3 Epstein markets (AOC, Sanders, Barak) that showed highest ILS in the pilot (0.55–0.93). Key findings: - AOC/Sanders high ILS is a formula edge effect (p_open≥0.94, denominator≤0.06) - Barak (17%→63%) shows genuine price discovery with anomalous Dec 20 crash - One wallet (0x4bfb41d5, veteran 5115 mkts) dominated all 3 from day 1 - 4 wallets appear in all 3 markets; none are newly created Also adds TASK_02F_FINAL.md synthesizing all 3 phases. Phase 4 (LLM Tier 3) gated on explicit user approval. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.