Beyond the pick.
Understand the price.

Current week · America/Chicago

Week 8 · 2026-11-02 – 2026-11-08

Publication windows: college football Friday at 9 p.m. Central; NFL Saturday at 7 p.m. Central. Reports appear when documented research is available.

This week’s full report & results

Top 5 research — proposed V3.3 format

Future NFL and college football candidates rank by documented model win probability where available. Unverified candidates remain unranked by probability, in supplied research order. V3.2 qualification thresholds remain unchanged. V3.3 is a proposed research framework.

No future Top 5 candidates published yet.

Each future Alternative requires individual Pokee adversarial review of contract, lineup/injuries, statistical evidence, opposing arguments, price/settlement and source freshness. PASS requires independent source retrieval and corroboration; evidence-packet conversation alone is INCONCLUSIVE. Unreviewed candidates cannot claim independent verification.

October 24–26 research results

College Football alternatives: 0–3 (0.0%). NFL Sunday alternatives: 2–1 (66.7%).

Weekend alternatives: 2–4 (33.3%)

Monday experimental candidates: 3–2 (60.0%) — separate from V3.2 alternatives.

Mixed-category descriptive total: 5–6 (45.5%) across 11 outcomes. This combines six weekend alternatives with five experimental Monday candidates, not verified V3.2 model performance.

Understanding Our Picks

SYSTEM — Highest Model Qualification

Model Approved: YES. A fully validated selection with net probability edge ≥5.0 percentage points, reliability ≥0.75 and reliability-adjusted edge ≥3.75 percentage points. Mandatory Red Team, data-quality and market-execution checks must all pass.

TACTICAL — Qualified Model Opportunity

Model Approved: YES. A fully validated selection with net probability edge ≥4.0 percentage points, reliability ≥0.50 and reliability-adjusted edge ≥2.0 percentage points. All mandatory checks must pass, and the selection must not qualify as SYSTEM.

ALTERNATIVE — Next-Best Research Candidate

Model Approved: NO. A research-supported candidate that does not satisfy the full SYSTEM or TACTICAL requirements. Each candidate explains its interest and failed or incomplete approval checks. No verified positive edge is implied.

Reliability and reliability-adjusted edge are uncalibrated heuristics. Top 3 lists prioritize category and retain the original research order within each category. A PASS or unsupported idea is not automatically an ALTERNATIVE. Historical and indicative odds are labeled; unavailable values are not estimated.

Official V3.2 performance combines SYSTEM and TACTICAL only. ALTERNATIVE performance is separate and requires verified, timestamped selection and settlement records. Categories stay tied to original pregame records; an alternative cannot be promoted after its outcome is known.

Running pick statistics

Official V3.2 record (SYSTEM + TACTICAL): 0–0 · n=0 · 0 pending. Alternatives are excluded.

How the picks are made

The framework studies independent forecasts alongside current information and market prices. Each selection is checked against defined qualification thresholds.

Explore each week

Full pick explanations, original research and outcome records live on their weekly pages. Weeks run Monday–Sunday in Central time.