All Edges

Every Edge, Every Card

PropsBot BKFC publishes free model picks for every upcoming Bare Knuckle Fighting Championship card, surfacing only the bouts where our 14-factor engine disagrees with the sportsbook line by enough to call an edge. Every pick is logged before the opening bell to a public accuracy ledger (currently 73.1% lifetime, as of July 2026). Picks include the fighter, model win probability, market-implied probability, and edge in percentage points. Live odds are aggregated every 30 minutes from xbet.ag, BestFightOdds, Polymarket, and Kalshi.

What do Advantage, Edge, and Predictability mean? +

Advantage (0-100): our model's estimated win probability × 100. 67 means the model gives that fighter a 67% chance.

Edge: how far the model's probability is from the de-vigged market probability. +11 means the model thinks the fighter is 11 percentage points more likely to win than the market implies.

Predictability (0-100): how much to trust this specific pick. Blends data quality, signal strength, factor agreement. We filter below 40 so they don't pollute the board.

Full betting glossary →

All Picks (0)

⏳ Waiting on sportsbooks

No edges to publish yet — bkfc-fight-night-uruguay opens in 2 days.

Edges = model probability vs sportsbook line. Books usually post BKFC moneylines ~7-14 days before an event. The 30-min odds scraper auto-picks them up the moment they drop. In the meantime, our model's pre-market favorites per card are below — no value call yet, just who we'd favor on clean data.

How to read this page

Advantage (ADV): our model's estimated win probability for that fighter, 0-100. 65 means the model gives them a 65% chance.

Edge: how far our model's probability is from the de-vigged market probability. 15pt means the model thinks the fighter is 15 percentage points more likely to win than the market prices them. Higher = bigger potential value.

Predictability: how trustworthy this specific pick is, 0-100. Combines data quality, signal strength, and factor agreement. 70+ = strong; below 40 filtered out.

⚠ warning: low-confidence edges — real data gap or market knows something. Don't chase big edges on thin data.