StatsDetective 8 | 2024 World Series Probabilities

Where to begin... probably with the worst performance in recent memory out of a baseball team. The Yankees lost the World Series, end of story. The Dodgers did not win, they really didn't need to. By my calculation, the series should not have ended at 5 games. In fact, the series should have been 3-2 in favor of the Yanks. But how do I know this?? I'm glad you asked.

Most analysis and prediction that I do is based on information that is known prior to a game being played. Obviously, this information is more prone to variation and much less reliable for prediction. The main reason is that while past performance and future performance correlate, they are not perfect predictors. Even the best have a bad game. Typically, I am hoping to be 60% accurate and that is on a good day. For games that already happened, the variables are much more powerful predictors. If I know how many hits each side had, I can get up to 70%+ and if more variables are included, the sky is the limit.

I am essentially asking a computer to tell me the outcome of a game given all the game information, something a computer is very well suited for. I used purely offensive models so the hits, walks, and stolen bases from both teams, 6 variables total. I used more than 2,000 games to train the model, then fed it the info from the World Series.

The results are detailed in the video but they are staggering. The Yankees management and pitching really fumbled what should have been a very competitive series. I myself am a casual Dodgers fan but something about a Yankees win would have felt good, I'm disappointed.