Honestly, I was a little nervous to meet with Professor Mario after the SSAC’17 conference, since my change of heart on my interests (Freedom to Return-The Last Step of a Hero’s Journey). The biggest concern was how I was going to indicate to him that I was appreciative of his help, while also telling him I was thinking about not doing sports statistics anymore.
However, Professor Mario has a good way about making me not freak out about the ambiguous future, because of his usual calm demeanor. After detailing what went well and what didn’t go well from the conference, he convinced me that perhaps I was just recovering from being sick and tired in Boston alone, and also being figuratively sick of sports statistics.
I told him my doubts about wanting to write a research paper over more sports statistics, and as we discussed what ultimate goals were I realized what was in question wasn’t writing this research paper or not. It was whether or not I wanted to use this research paper as a means of better learning data analytics. We got into a discussion about having a toolbox, and how how every skill I acquire through learning experiences either sharpens a tool I already have or gives me a new one. At the end of the day, I want a toolbox that has a variety of sharp tools, not just a great sports statistics research paper. Therefore, I came to the conclusion that as long as I kept the “toolbox mentality” I didn’t have to worry too much about getting “sick” of sports statistics, because it is just a means of enabling me to learn more about data analytics…
Digging further, I came up with a new idea on a question to measure.
As the NBA season has been wrapping up there has been a lot of talk about who should be MVP…one argument that is always brought up is whether or not a player should be regarded as “less good” because they are a high scoring with a bad team…
For example, D. Booker, a 19 y/o scored 70 points in a game, but his team still lost
A. Davis has had multiple 50 point + games, but his team still looses
R. Westbrook suffers the same phenomenon…
So my question is, “What factor, if any, should we “scale” high scoring players on bad teams, in trying to object evaluate the amount of points they score compared to equally high scoring players on good teams?” (0-1)
Good/Bad team= determined by season record
For this argument, we can look at these factors/stats:
- It’s easier to score when you’re on a bad team because you get the ball more: touches relative to average touches in their position (PG to PG)
- It’s easier to score when you’re on a bad team because defense is more relaxed: defensive sloppiness factor
Against this argument, we can look at these factors/stats:
- It’s easier to score when you’re NOT on a bad team because you are more likely to have sloppy plays: turnover rate, points off turnovers (execution)
- It’s easier when you’re NOT on a bad team because you are more likely to have defense’s attention divided because multiple people on your team are a threat: defensive sloppiness factor
Based on these two factors I can use old-fashioned statistics to figure out the weight of each of these factors and do pluses for FOR and minuses and AGAINST and calculate a factor to multiply to the score of players to scale it objectively.
I also need to figure out a scale of good to bad teams so that this factor can be multiplied to all players so that’s a part to expand on…
Overall I’m excited to have something to work on again, and even this second brainstorm around I feel as if my brain is more agile than it was during the fall brainstorm. I definitely feel like I’m riding the learning curve of thinking of how to start a data analytics problem!
Hopefully this opens more doors in terms of helping me identify what I can grasp quickly and what I can’t so as I search for the best full-time job my senior year next fall, I will be able to best pitch myself!