SIGNAL DASHBOARD · MAY 2026
Quantitative Prediction Market Signals
ALL-TIME ACCURACY
92.3%
12/13 resolved
BRIER SKILL SCORE
+0.49
vs 0.0 naive baseline
PEAK SIGNAL EDGE
+53pp
Chimaev 61% vs 8¢ market
WC 2026 SIGNALS
22
pre-logged · opens ~Jun 1
OPEN SIGNALS6 active
| DATE | PREDICTION | DOMAIN | MODEL | MARKET | EDGE | RESOLVES |
|---|---|---|---|---|---|---|
| 2026-05-03 | Khamzat Chimaev wins UFC 328 MW Title | UFC | 61.3% | 8¢ | +53.3pp | May 9 |
| 2026-05-03 | Tatsuro Taira wins Flyweight Title | UFC | 84.1% | 62¢ | +22.1pp | May 9 |
| 2026-05-03 | Volkov beats Cortes-Acosta — UFC HW | UFC | 73.8% | 63¢ | +10.8pp | May 9 |
| 2026-05-03 | LOUD beats Cloud9 — VCT Americas | ESPORTS | 61.2% | 61¢ | today | May 3 |
| 2026-04-12 | Arsenal advances UCL semi-final | FOOTBALL | 58.7% | 52¢ | +6.7pp | May 5 |
| 2026-04-15 | Psychedelic therapy FDA approval 2026 | HEALTH | 54.0% | 47¢ | +7.0pp | Dec 31 |
RESOLVED PREDICTIONS12/13 correct
| PREDICTION | EDGE | RESULT |
|---|---|---|
| Raptors win G6 · NBA | +15pp | CORRECT |
| Pistons win G6 · NBA | edge | CORRECT |
| Rockets win G6 · NBA | +21pp | MISS |
| Q3 2025 Esports · 10 predictions | avg +20pp | 10/10 |
WORLD CUP 202622 signals · opens ~Jun 1
| MATCH | MODEL | EST. EDGE |
|---|---|---|
| Argentina vs Haiti | 97.3% | +47pp |
| Spain vs Haiti | 95.4% | +45pp |
| Brazil vs Jordan | 94.5% | +45pp |
| France vs Haiti | 94.3% | +44pp |
| + 18 further predictions pre-logged | ||
MODEL COVERAGE8 integrated models · real-time WebSocket detection
UFC/MMA
42
fighters · 62 active markets
TENNIS
27
ATP/WTA · surface-adjusted
NBA + SERIES
16
teams · game + series
ESPORTS
73
teams · CS2, Val, LoL
MLB
30
teams · daily signals
CRICKET
20
ICC T20I nations
PEAD EARNINGS
8K+
US companies
GET THE SIGNALS
from £75/month →
◈
METHODOLOGY
Bradley-Terry Elo (M6 joint logistic, Li & Li 2025) · Glicko uncertainty intervals · Price velocity signal (M5) · Google Trends 2D-IPW narrative correction · Rolling form-weighted Elo · Platt calibration auto-updating on every resolved prediction. All predictions timestamped publicly before events — no look-ahead bias.