20180926, Basketball Stats ... by on Scribd

2018-2019 NBA Fantasy Rankings

Greetings from the back of my spreadsheets. I am providing to you, my fellow fantasy basketball enthusiast a glimpse into what I do for a hobby: play fantasy basketball. I am 18 year veteran of fantasy basketball and the winner, and loser, of several leagues. Whether playing in casual or competitive leagues if there is one thing that I love to do is to watch my players succeed in the virtual arena. Similar to the stock market, I would hit refresh on my eight opened ESPN/Yahoo sports tabs following my players every play-by-play hoping that they would rebound, score, assist, or steal their way to a great game. Sometimes players had good games, sometimes they had bad games but what I soon learned is that the players all followed a similar trend that with a little bit of investigation I could figure out. I call this investigation data analysis, or stat crunching.

​As a result of trying to gain a competitive edge I started to use expert’s picks and usually had mixed results. I started noticing that sometimes the experts would pick a player because they would grace ESPN’s TOP 10 highlights or would be a popular or well-known picks. It seemed to me that this was a major criteria for their selections. Therefore, in the pursuit of better data driven decision I decided to formulate my own rankings, based on the vast knowledge that my Stats and Calculus professors required me to learn. 

​I started off by using a technique that a majority of the experts use now to rank players, the “Z-score”. The z-score, is a way to standardizing a NBA basketball player’s statistics by fitting all the players’ statistics against a normal distribution, or bell shaped curve. Analysts will take all player’s statistics and scale the data down so that the average player will have a score of close to 0. In this manner they take all of their sortable statistics and standardize them so that they can compare them together and against each other. Just think about how you would compare a player’s points scored per game (PPG) compared against their Free Throw Shooting Percentage (FT %). With that in mind, these scores are then figured out for each category and then added together, with whomever had the highest score would be ranked higher. Overall it is an intuitive way of ranking that majority of basketball stats junkies understand. Unfortunately it leaves a lot to be desired as it does not consider some of the fundamental lessons that we have learned throughout the years of playing. Ultimately some stats are scarcer than others. How do we account for this amongst other things?

​What I started to use three years ago is a method of ranking that is part of the platform of Multi Criteria Decision Making (MCDM) methods named the Technique for Order Preferenceby Similarity to an Ideal Solution, or TOPSIS. TOPSIS is a mathematical way of doing literally its name describes, rank by a predetermined preference. The best choice, or ultimately the best ranking, should have the shortest distance to the most positive ideal solution and be the furthest away from the least ideal, or negative, solution. This model allows for a user to select a weighing criteria, determined by scarcity of statistics, and find the solutions that are closest to the ideal solution, or how the ideal player should perform. Think Andre Drummond’s rebounding, mixed with Draymond’s stealing, a little bit of Rudy Gobert’s blocking, and James Harden’s overall production without the turnovers. That is what TOPSIS aims to lead you to. 

​The strategy of using TOPSIS is to rank players based on these “super-player” attributes and select them in order to find the players that have the statistics that will most help and least hurt your team. Some of the picks are flashy. Some of them you won’t agree with. But they do provide insight into building the foundation of your team (principally picks #1-#10) before you have to worry about specialists. So without any more fanfare, though I plan to describe this process more in detail in the notesof my blog, I present my rankings. If you are uneasy just believing me I have included their z-scores as well and have taken the liberty of averaging them against these rankings to provide a “consensus” ranking similar to what FantasyPros does. Enjoy.