Live Win Expectancy — In-Game Model

Pre-game models tell you who should win. In-game models tell you the live fair price after every pitch. Built on a Monte Carlo simulator that integrates the base-out run-expectancy matrix. This is how live betting markets should be priced — and aren't.

--Current Home Win %

--Pre-Game Home %

--Win Prob Swing

--Games In Progress

--Last Poll (UTC)

1,000MC Per State

📊 Win Expectancy Throughout the GameDodgers @ Padres

Probability that LAD wins, recomputed after each event. The shaded area is the home team's lead. Sharp dips/spikes mark high-leverage plays — those are the spots where live betting markets are slowest to react.

📋 Play-by-Play Timeline

Inning	State	P(home)	Fair Home	Fair Away	Event

🔢 Base-Out Run Expectancy (RE24)

Bases	0 outs	1 out	2 outs

Expected runs scored in the remainder of the inning from each (bases, outs) state. Empirically derived from millions of historical PAs. The bedrock of every live-betting model.

🧠 How the live model works

After every pitch, the game state is:
   (inning, top/bottom, outs, bases, score_diff)

For example, after Ohtani's first-inning HR:
   inning=1, top=False, outs=0, bases=(0,0,0), score_diff=+1

The model takes that state and runs 1,000 forward simulations:
  1. Apply the RE24 table to project remaining runs this half-inning
  2. Sample Poisson runs for each remaining half-inning
  3. Handle walk-offs, extras, and ties
  4. Count: home wins / total

That ratio is P(home wins | current state).

Convert to American odds, add target vig, and you have the *fair* live
moneyline.  Compare it to what the book is offering — every gap is a
potential live bet.

Sharpest live-betting edges:
  • After a HR — books overreact, fade the moving line
  • Bases loaded, 2 outs — books underprice the strikeout probability
  • Late-inning ties — bookies regress; you trust the better bullpen
  • Pitcher fatigue past 90 pitches — TTO penalty is bigger than books model

⚡ Live Edge Detector

Every 60 seconds, we compare our live fair price to the book's live moneyline across all six tracked sportsbooks. Any gap larger than 4% triggers an alert. Live markets are slower and squarer than pre-game markets — that's where the in-game edges live.

Pre-game: Books re-shade on 6-figure handle and sharp money. Hard to beat.
Live: Books re-shade on a fixed algorithm that updates after each pitch. The algorithm is sometimes wrong by 3-5% in high-leverage spots.
Late innings: Books often run a vig of 8-12% on live moneylines. You need a real edge to overcome that.