In recent years, advanced metrics have dominated the game of Major League Baseball. One such metric, the Pythagorean Win-Loss Formula, is commonly used by fans, reporters, analysts and teams alike to use a team’s runs scored and runs allowed to estimate their expected winning percentage. However, this method is not perfect, and shows notable room for improvement. One such area that could be improved is its ability to be affected drastically by a single blowout game, a game in which one team significantly outscores their opponent.<br/>We hypothesize that meaningless runs scored in blowouts are harming the predictive power of Pythagorean Win-Loss and similar win expectancy statistics such as the Linear Formula for Baseball and BaseRuns. We developed a win probability-based cutoff approach that tallied the score of each game once a certain win probability threshold was passed, effectively removing those meaningless runs from a team’s season-long runs scored and runs allowed totals. These truncated totals were then inserted into the Pythagorean Win-Loss and Linear Formulas and tested against the base models.<br/>The preliminary results show that, while certain runs are more meaningful than others depending on the situation in which they are scored, the base models more accurately predicted future record than our truncated versions. For now, there is not enough evidence to either confirm or reject our hypothesis. In this paper, we suggest several potential improvement strategies for the results.<br/>At the end, we address how these results speak to the importance of responsibility and restraint when using advanced statistics within reporting.
For years, Commissioner Rob Manfred has hinted and brought about the idea of adding two more teams to Major League Baseball (Mitchell). The growth of the game is of utmost importance, and they have made many changes to try to expand the growth of fans the past few years particularly catered to new and young fans. New rules like a pitch clock and mound visit limitations are examples of in game changes made to speed up the game, but they have also experimented with spring training and regular season games internationally or at new venues. In just the past decade, games have been played or planned (due to COVID-19 cancellations) in Monterrey, Mexico City, London, Tokyo, San Juan, Montreal, Las Vegas, Williamsport, and even Iowa. With the exception of the Williamsport Little League Classic and the Field of Dreams game in Iowa, all these locations had games to see what the atmosphere and logistics would be like with expansion in mind as a possibility in the future. With this in mind, this thesis will analyze and come to a conclusion on the following cities for the best fits for expansion: Monterrey, Mexico City, San Juan, Vancouver, Montreal, Las Vegas, Portland, Nashville, Raleigh, and San Antonio.
This thesis aims to develop a new way to value players for all teams in the MLB, despite the financial disparity. Displayed in the rest of this paper, is a player valuation model created around each team's salary level, focusing on the player’s offensive output. The model functions in a way that values players by their ability to help their team score runs and win games by setting parameters for salary expectations based on player performance. This allows for small market MLB teams, like the Cleveland Guardians, to build a roster of players around their specific salary limit, specifically to score the maximum runs and win games. On the contrary, the model also works for big market teams, like the Los Angeles Dodger, allowing them to project their larger salary limit to players and build their ideal roster as well.