Postseason
Two nested simulations: finish the regular season, seed both leagues, then play the bracket twenty thousand times. Every number on this page is a probability, and the most important ones are the small ones.
A short series barely settles anything, and that is not a weakness of this model but of the format. Below is the exact probability that the better club wins, given a flat per-game edge. It is arithmetic, not simulation.
The best team in baseball is roughly a 57% favourite in a typical game. Such a team loses a best-of-five more than a third of the time. Any postseason forecast that sounds confident is overselling.
Two independent reads on team strength, side by side rather than averaged, because where they disagree is more informative than a blend. Elo is the continuously updated rating that tested best at game level. Pythagenpat is expected win percentage from this season's runs scored and allowed.
Chance to win the World Series under each strength estimate. The grey bar between the dots is the disagreement, and it is a floor on how uncertain every number on this page is.
Distribution of games played, across all simulations.
Every matchup below is a long shot. That is the point.
Since 2022 the three division winners take seeds one through three regardless of record, and the wild cards take four through six. A weak division winner therefore gets a bye and home field over a much better wild card. The seed column in the table above is each club's most frequent seed across the simulations, and the quirk is visible in it: Cleveland at 84 wins is seeded second in the American League and takes a bye, while the 93-win Yankees open on the road as the fourth seed.