Bottling It, Or Bad Math?

Bottling It, Or Bad Math?

The shift of the 2024 World Series to Yankee Stadium for Game 3 did nothing to impede the momentum of the Los Angeles Dodgers. Freddie Freeman sparked the Dodgers early by hitting a home run for the third consecutive game, while Walker Buehler, along with the Dodgers’ bullpen, effectively silenced the New York Yankees’ offence once again.

Heading into this series, the Yankees were touted as having a significant advantage in starting pitching. However, the Dodgers have decisively turned that expectation on its head. In Game 3, Buehler delivered a masterful performance, pitching five scoreless innings. This followed impressive outings in the first two games by Yoshinobu Yamamoto, who allowed just one run in 6? innings during Game 2, and Jack Flaherty, who gave up only two runs over 5? innings in Game 1. Collectively, the trio has posted a remarkable 1.62 ERA, the lowest recorded by a team’s first three starters in a World Series since the Cleveland Indians in 2016.

Among the three, Buehler’s performance was perhaps the most unexpected. He made his return from a second Tommy John surgery midway through the season, struggled with a 5.38 ERA over 16 regular-season starts, and appeared to have lost some of the power in his fastball. Yet, he rose to the occasion when it mattered most. Historically, when a team takes a 3-0 lead in a best-of-seven series, they have done so 40 times, with 31 of those instances resulting in sweeps. The only team to have ever overcome a 3-0 deficit in such a series was the 2004 Boston Red Sox, who did so against the Yankees.

This raises the question: have the Yankees choked in the World Series, or was their performance simply not as strong as anticipated?

Before the 2024 World Series, the Los Angeles Dodgers were viewed as slight favourites with betting odds of -120, implying an approximate probability of 54.55% to win the championship. This confidence reflected their strong regular-season performance and recent success in the playoffs. Meanwhile, the New York Yankees entered the series with odds of +110, suggesting an implied probability of about 47.62% to clinch the title. These odds highlighted the competitive nature of the matchup, with the Dodgers holding a marginal edge over the Yankees as they prepared to face off in the World Series.

To simulate the World Series matchup between the Los Angeles Dodgers and the New York Yankees, we employed a systematic approach that combined performance metrics with statistical modelling. The initial step involved adjusting the teams' base scores, which serve as key performance indicators encompassing areas such as hitting, pitching, defence, and speed. These base scores provide the foundation for our simulation. To ensure the analysis was relevant to current conditions, we incorporated each team's recent performance over the last 30 games. By calculating the win ratio and applying a multiplier to the base scores, we effectively ensured that teams demonstrating strong recent form would have their capabilities accurately reflected in the simulations.

In our simulation methodology for the World Series matchup between the Los Angeles Dodgers and New York Yankees, we employed a normal distribution model to generate scores for both teams. This model utilised each team's adjusted performance metrics, which included offensive and pitching statistics, to establish expected scoring outcomes. For the home team, we applied a slight home-field advantage factor, increasing their expected score to reflect the reality that teams often perform better in familiar environments. The scores for each team were generated by sampling from these normal distributions, ensuring that the results captured the inherent variability in sports performances. This approach allowed us to accurately simulate individual games while incorporating key competitive dynamics, ultimately providing a realistic estimation of the series outcomes.

Our simulation rigorously adhered to the traditional best-of-seven playoff format, closely mirroring the actual structure of the World Series. Each team was required to win four games to secure the series victory. To replicate the real-life scheduling of playoff series accurately, the order of the games alternated between the two teams, ensuring that home and away contexts were faithfully represented. This design choice ensured that our simulation captured the intricacies of team dynamics and the competitive nature typical of a World Series matchup.

To estimate the winning probabilities for each team, we employed a Monte Carlo simulation technique, executing one million iterations of the best-of-seven series. This substantial sample size provided a robust statistical foundation for our analysis, yielding significant data from which we could derive precise estimates of each team's likelihood of success. Throughout the simulation iterations, we systematically tracked the cumulative number of series wins for both teams. The winning probabilities were calculated by dividing the total number of wins for each team by the total number of simulations conducted. This meticulous approach allowed us to generate a comprehensive understanding of the potential outcomes in the World Series matchup, effectively quantifying the impact of performance metrics and game dynamics on overall series results.

The simulation results indicated a competitive edge for the Los Angeles Dodgers in the hypothetical World Series matchup against the New York Yankees, with the Dodgers having a probability of winning the series at approximately 52.94%, while the Yankees held a 47.06% chance of securing the championship. This marginal advantage suggests that, based on the performance metrics and home-field dynamics factored into the simulation, the Dodgers were slightly more likely to emerge victorious.


In terms of actual series wins derived from the one million iterations, the Dodgers triumphed in 529,399 simulated series, whereas the Yankees won 470,601 times. This outcome reflects the Dodgers' statistical advantage and underscores the competitiveness of the matchup, with both teams demonstrating strong potential throughout the simulations. The distribution of victories highlights the tightly contested nature of this hypothetical series, suggesting that while the Dodgers were favoured, the Yankees were not far behind and could easily clinch the title under varying game conditions.

These simulation results closely align with the implied gambling odds preceding the World Series. The betting odds indicated that the Dodgers were favoured at -120, which translates to an approximate 54.55% chance of winning, while the Yankees were listed at +110, implying a 47.62% chance of victory.

The simulation’s output of 52.94% for the Dodgers is quite close to the betting line, reinforcing the idea that both analytical modelling and public sentiment converge around similar assessments of the teams' strengths.


As the New York Yankees find themselves trailing 3-0 in the World Series against the Los Angeles Dodgers, the hand of history is decidedly unfavourable to them. My simulation model, which closely mirrored the betting odds, suggested a competitive matchup, reflecting expectations of a tightly contested series. However, recent performance has not aligned with these expectations. Star player Aaron Judge has struggled to find his form, while Juan Soto's uncertain contract situation looms large over the team's morale. In stark contrast, the Dodgers' Freddie Freeman is on the cusp of setting a record in the World Series, showcasing his exceptional performance when it matters most.

This combination of factors raises serious concerns about the Yankees' ability to mount a comeback. The statistical landscape hinted at a closely matched series, yet the reality is far from it, leaving little to suggest that the Yankees can avoid what appears to be a significant choke. Historically, only one team, the 2004 Boston Red Sox, has managed to recover from a 3-0 deficit in a best-of-seven series. Barring a historic turnaround, it seems the Yankees are on the verge of falling short of expectations in what was anticipated to be a fiercely competitive World Series. The math might suggest possibilities, but the reality on the field tells a different story.

Lee Grant

Certified Azure Data Engineer. Helping customers modernize their data for Azure and other public cloud platforms.

3 周

As a data engineer (and longtime Dodger fan) I really enjoyed this analysis. It may just come down to intangibles that cannot be modeled accurately. Aaron Judge is in his head and most of the rest of the team cannot pick up the slack. Ohtani is having a rough postseason, but the Dodgers are getting production from deep in the order (Edman, K. Hernandez). Great article!

要查看或添加评论,请登录

社区洞察

其他会员也浏览了