The most valuable player is an important criterion to see how well a player is. The most famous players in NBA history may also reward this title during their career. For many years, the MVP evaluation process is hard to determine the first and second place because there is no exact rubric for Journalists and commentators to make their judgments and the process is more controversial at all. In our project, our purpose is to build statistical models to validate who should win the 2015-2019 regular seasons’ NBA Most Valuable Player (MVP) Award. Prior to building the MVP models, the player statistics data has been Z-standardized to make sure the equal weight of each category and remove any mean and standard deviation bias. Our second model, the “Uniform MVP Index” has been derived from combining each player’s Z-statistics in every category with equal weight. The Uniform model is a very simple proto-model to demonstrate and define the MVP Index. Then, teams have further derived a "Weighted and Subset" model by adding the weight factor and the best subset feature selection. Doing a comparison among the weighted model and best subset model, the team further agrees to prefer the subset model based on inflated power multicollinearity. Authors have added the "Team Winning'' factor in the Power Model from power= 0 (equivalent to the Weighted Model), 0.5,1,1.5 to power= infinity (MVP choosing from the best Team). The Power MVP Index will be multiplied by the power of the team winning% in the Power model. In this model, we will see how the player's performance could affect their team's results (winning%). Based on the Validation of the 2015-2019 Seasons, there is no single MVP model that could determine the MVP winner from year to year, but in fact, we could set a high limit of power equal to 3. Thus our model could be more consistent in predicting MVP than the NBA official EFF index model.
0 Comments
By building statistical models, this project is to show how age is a big factor for NBA players. Using the trend for former players to predict the active NBA player's career future. 36 valid candidates have been chosen who completed at least 1200 games and 15 seasons (excluding the short season or players who did not complete more than 20 games) in their career. Doing Z-Standardization and removing any standard deviation, means bias before creating each age model. Combining the Z statistics of each player to make sure equal weights for players' Career Average. The purpose of this project is to see how age affects a player's total playing time and total points received in each season add on. To predict active duty NBA players' career future, authors calculated the combo average of three categories (point average per game PPG, minutes played per game MPG, and points per minute PPM) with 36 sets of data and concluded the golden age range for players. By using statistical analysis software JMP and Minitab, the authors partitioned 36 players into 7 clusters, and the multi correlation has been studied. Contrasting the correlation of players having the same position in each cluster, to see if the value is close to their cluster average, the project is to show how two players are similar to each other and to predict out the active duty player’s future performance by using the former player career trajectory.
|
AuthorCategories |