2026 playoff odds

One plate appearance at a time. For each batter against each pitcher, the simulator works out the chance of a strikeout, walk, single, double, triple, home run or out. It starts from each player's own rates. It then adjusts for the league, for whether the batter and pitcher throw and hit from the same side, for the ballpark, and for home field (home teams strike out less and walk more; equal teams win about 53% at home, as in real games). Runners move the way they really moved in 2025, given the outs, the bases and what the batter did.

Player talent. Each player's rates blend his 2024, 2025 and 2026 regular seasons, with recent years counting more. They are then pulled toward league average, more for players with little track record. A pitcher's luck on balls in play is mostly ignored. His team's fielding decides how many of those balls fall for hits.

Lineups. Managers mostly repeat themselves, so the model builds each lineup from the club's recent ones. A start counts for more if it came in the postseason, came recently, or came against a starter who throws with the same hand as tomorrow's. The nine hitters with the most weight play, in their usual order.

Starters and bullpens. An announced starter starts. Otherwise each club cycles through its postseason starters, and a pitcher needs four days of rest unless nobody has had them. Starters leave after facing batters, the 2026 postseason average. The rest of the staff pitches as one bullpen, weighted toward the relievers who pitched most this year.

What it leaves out. Pinch hitters, defensive swaps, a manager matching relievers to hitters, injuries during a series, and fatigue. Young players are not projected to improve, so a breakout season counts only partly.

A check against the regular season.