Statistical Analytics · Research-Validated · No Predictive Claims
UK49s Statistical Analytics Platform
Scoring Method
Descriptive v1
Section 1 — System Health
Section 2 — Statistical Summary
Descriptive Factors (not predictive)
Section 3 — Draw Overview
| Draw | Data Maturity Score | History | Status | Hit Rate 2026 | Baseline | Trend |
|---|
Section 4 — Top 5 Statistical Rankings · Section 5 — Score Breakdown · Section 6 — Score Gap · Section 7 — Method Performance (validated null result)
Section 8 — Performance Tracker
Section 9 — Sample Size & Hit Rate Monitor
ℹ️ Brunchtime/Drivetime have under 150 verified draws -- treat their rates as low-confidence until more data accumulates.
⚠️
Score Separation Notice
Score separation between rank 3 and rank 4 is narrow -- this reflects how close the descriptive scores are, not a predictive signal.
Section 11 — Daily Automation Status
Section 12 — Statistical Notes
Scoring method: Descriptive v1 (no validated predictive edge -- see Section 13)
Section 13 — Experiment Tracker
| Experiment | Status | Result | Notes |
|---|
| Historical data integrity | Contaminated -> Fixed | 95%+ of original history | 2026-06-14 bulk import found fabricated; replaced with verified scrape of teatimeresults.info |
| REPEAT | Tested — No Edge | 12.3% | Baseline 12.2% -- re-validated on clean 7-8yr data, train/test split |
| MULTI | Tested — No Edge | 11.4-11.5% | Baseline 12.2% -- below baseline on both LT and TT |
| T1 Mirror | Tested — No Edge | 12.2-12.5% | Baseline 12.2% -- fixed-offset mapping, no measurable lift |
| Booster Calls | Tested — No Edge | 11.8-11.9% | Baseline 12.2% -- no measurable lift on clean data |
| Rolling Hot/Cold | Tested — No Edge | 12.1-12.5% | Tested 4 window sizes (20/50/100/200) on clean data, all at baseline |
| Cross-Draw (LT→TT same day) | Tested — No Edge | 0.761 vs 0.735 exp. | No meaningful difference from independence |
| Pairwise Co-occurrence | Tested — No Edge | 0/15 replicated | Train/test split across 1,176 possible pairs, none survived |
| Markov Transitions (all pairs) | Tested — No Edge | 0/20 replicated | 2,352 possible A→B pairs tested, regression-to-baseline pattern |
| Uploaded strategy guides (chase rules, date arithmetic, etc.) | Tested — No Edge | At or below baseline | 2 strategies mathematically proven to be disguised doubling (zero information content) |
| Coverage Builder | Planned | — | Budget-spread tool -- does not claim to improve odds, only changes variance |
| Anomaly Monitor (automated) | Planned | — | Scheduled version of the integrity checks that caught the contamination above |