free sign in →
← back to pipeline
mr

Will Choo Kyung-ho win the 2026 Daegu mayoral election?

will-choo-kyung-ho-win-the-2026-daegu-mayoral-election · Daegu Mayoral Election Winner · category: Elections, Global Elections, Main Election, Politics, South Korea · resolves
run_id: 61588d35-3faf-4643-b996-e5df48b7d7ed · started · status skipped · planner_skipped · ✕ planner rejected
no_trade: action=OMIT
signal detected

detect_reversal

● executed
direction
DOWN
current / peak
74.5% / 76.5%
move_start
67.5%
reversal
+2.0pp
move size / hours
9.0pp / 22.0h
shape
steady
72h range
64.0%—76.5% (Δ12.5pp)
since peak
1.5h

Price rose 9.0pp from 67.5% to 76.5% over 22.0h (steady), then reversed. Now at 74.5%, down 2.0pp from peak, 1.5h after the peak.

news gathered

news retrieval (tomographic)

● executed
raw→filtered: 100 → 1 quality: HIGH tokens: 4571
AI verdict

Prompt-MR · misprice analysis

● executed
OVERPRICED conviction: LOW opportunity: SMALL action: OMIT
gpt-5 · prompt_2_misprice_v9.0_theta_v4 · 12995 tokens
Semantic Polarity: YES=Choo Kyung-ho wins Daegu mayoral race; NO=he loses. Trend is UP, pricing in the YES-World. Momentum check: E=0, D=0, L=0 (R=0 unverified). Incremental only: E=0, D=0 (priced-in: E=0, D=1). Self-check: no change. Theta direction: NEUTRAL; Price move direction: UP (from 68% to 76%); Alignment: N/A; Theta: NEUTRAL — election market (no time-decay). No ★ DIRECT resolution evidence (e.g., new reputable poll showing Choo ahead, endorsements materially shifting odds, or official results) appeared during the move; the only priced-in context points to a razor-thin race, which does not justify a 9pp climb. Given the CIRCUIT BREAKER — STALE NEWS warning (feed likely missed the catalyst), I use LOW conviction and SMALL size, but on captured evidence the move looks partially unsupported and prone to some mean reversion.
rule audit

audit_gate

● executed
ALL
trade plan

position_planner

○ skipped
✕ no_trade: action=OMIT

the deterministic planner rejected this trade. typical causes: poor R:R, spread exceeds gap, settlement danger zone, market too close to resolution. LLM assessment is above (Stage 6); planner is a pure-math gate layered on top.

step post-omit

news watch (A.2 / A.3)

● executed
watch: 5f691ad7
status: ⏳ expired
armed: 2026-05-13T05:13Z
closed: 2026-05-14T05:34Z
polls: 5
signals seen: 4126
decision price: 74.5pct