The Story Behind The Model

Hello and Welcome! This model is built by Jake Samson, a Business Analytics Masters student at Drexel University, with a passion in sports statistics. Every set of power rankings is different, some based on statistics, some based on the eye test, and many in a combination of both.

This is a strict ten-variable analytical model, based on ten statistics with different weights attached based on importance (they will be listed below shortly). After normalizing the data and creating Z-Scores (scores above or below the average of each statistic), the teams will be ranked from 1-30 with a total overall Z-Score. For example, if the Mavericks have a .74 Z Score in one statistic, and a -.34 Z Score in another statistic, and each statistic has the same weight, their overall Z Score puts them at a .4.

The database will be accessible via Google Sheets at this link here: https://docs.google.com/spreadsheets/d/1NI0_WYjrDAcrMZCnWfn-RRHEPfpkeFsYdYf0citX53M/edit?usp=sharing. The sheet will be updated and a new power ranking write up will be released on Mondays. Without further ado, the ten statistics that this model will be based off of with descriptions.

Offensive Rating (25%): The top teams have a top offense, it's just that simple. 14 of the 16 playoff teams ranked in the top 16 when it comes to Offensive Rating in 2021-2022. There is a large formula for this attached, but lets not clog up the intro page.

Three-Point Percentage (15%): In the world of sports analytics, the three pointer has been the rising star in the NBA. The best teams shoot a lot of threes, and make a lot of threes with the average of 35 three point attempts per game last season.

Defensive Rating (10%): The saying goes "good offense beats good defense" and the same goes in this ranking. The correlation is not as strong for playoff teams for this ranking, but it's still important. See the formula attached here.

Pace (10%): Pace is the number of possessions per 48 minutes. In the NBA this season, the average Points per Possession is 1.11, so in a perfect world, take that and multiply it by the Pace to equal the total points per game.

Turnover Differential (10%): A turnover is an empty possession, and gives the other team one often in transition.

Free-Throw Attempt Differential (7.5%): Good teams have good players who get fouled. They will make teams pay based on the number of free throws they get.

Rebounding Differential (7.5%): A huge part of the game, the importance of cleaning the glass can be a momentum shifter that goes even beyond analytics.

Margin of Victory (5%): Can also be the margin of defeat, depending on how the team is looking.

Strength of Schedule (5%): If you've played good teams, it can give a slight bump as you've competed against the best.

Assists on Field Goals (5%): A stat that is often overlooked. There are some tremendous players like Luka Doncic or Stephen Curry who can make these unassisted plays look easy, but most good teams can move the ball well. Despite the low weight, it definitely deserves a shoutout in this model.