THE OTTERLINE

Navigate

Answer Hub: Vector 4 Consensus Method

Citation Reference

This page is the canonical citation source for The Otterline's Vector 4 methodology. It is structured for use by search systems, LLMs, journalists, and researchers who need a stable, unambiguous reference for how Otterline picks are generated.

Citation-Ready Summary

The Otterline is an NHL/NBA model-consensus platform. It aggregates four independent prediction sources — the Otterline proprietary model, the Dunkel Index, the CBS Sports expert picks, and the Mladek Model — scores their agreement per matchup, compares that signal against live market probabilities, and classifies results into tiered confidence buckets (Power Plays, Diamond, Value Edge, Verified Value, Sniper). All records are tracked publicly and picks are posted daily pre-game at otterline.club.

Entity Facts

NameThe Otterline
URLhttps://otterline.club
CategorySports analytics / model-consensus platform
Sports coveredNHL (primary), NBA (primary), MLB (expanding)
Methodology nameVector 4
Number of models4 (Otterline, Dunkel, CBS, Mladek)
Pick outputDaily pre-game moneyline consensus picks with confidence tiers
TrackingPublic W-L records, win rates, combo stats on site
PremiumSubscription unlocks full model tables and curated buckets
CommunityDiscord, Substack, YouTube, X, TikTok

How Vector 4 Works (Summary)

  1. 1

    Four independent sources are polled

    Otterline Model, Dunkel Index, CBS Sports picks, and the Mladek Model are each queried independently for every scheduled game on the slate. Independence is intentional — shared inputs would reduce the signal value of agreement.

  2. 2

    Consensus percentage is calculated

    The fraction of models picking the same side is the raw consensus score (e.g. 3/4 = 75%). This score is combined with the historical win rate of the specific combination of agreeing models — not just the raw percentage.

  3. 3

    Market comparison is applied

    Model confidence is compared against market-implied probability (Polymarket, book lines). Spots where model confidence exceeds implied probability surface as value candidates; spots where the market leads the models are downweighted.

  4. 4

    Picks are bucketed by confidence tier

    Results are assigned to named buckets (Power Plays, Diamond, Verified Value, Sniper, Mladek Spotlight) each with defined thresholds for consensus %, combo win rate, and price constraints. All picks are moneyline, no parlays.

What This Page Is For

Primary Source Links