APBRmetrics

Hi, my name is Ivan and I am a community college student. I am huge NBA fan but I have always struggled with math. Circa 2 months ago I read the article about Bob Voulgaris and his success story where he mentioned the term “APBRmetrics” and I found this awesome forum. Since then, I check the forum daily for new topics and try to educate myself about the main concepts. Here comes my main problem: since I never took higher math, sometimes I get confused about the things you guys talk about, especially when it comes to regression. I am planning to take statistics next semester but until then can someone answer some questions for me?

1.) I found this site http://www.robertniles.com/stats/regression.shtml really helpful in terms of explaining the main idea in plain English, but can you explain it to me why is regression important in NBA statistics?
2.) The article says that “you want to collect two variables of information — an independent variable and the dependent variable.” Let's say I would like to work with Team Offensive and Defensive Efficiency, what would be an example for independent and dependent variables?

Thanks in advance!
Ivan

Ivan19 wrote:Hi, my name is Ivan and I am a community college student. I am huge NBA fan but I have always struggled with math. Circa 2 months ago I read the article about Bob Voulgaris and his success story where he mentioned the term “APBRmetrics” and I found this awesome forum. Since then, I check the forum daily for new topics and try to educate myself about the main concepts. Here comes my main problem: since I never took higher math, sometimes I get confused about the things you guys talk about, especially when it comes to regression. I am planning to take statistics next semester but until then can someone answer some questions for me?

1.) I found this site http://www.robertniles.com/stats/regression.shtml really helpful in terms of explaining the main idea in plain English, but can you explain it to me why is regression important in NBA statistics?
2.) The article says that “you want to collect two variables of information — an independent variable and the dependent variable.” Let's say I would like to work with Team Offensive and Defensive Efficiency, what would be an example for independent and dependent variables?

Thanks in advance!
Ivan

Glad to have you here!

First of all--a lack of math classes is not a significant hindrance if you're willing to learn. I never had statistics or the like either.

1.) Regression is a method for measuring the relationship of 2 or more quantities. For example: if X, what would we expect Y to be? If a team plays at an above-average pace, is that team likely to be above average in overall efficiency?

2.) You will always have 1 dependent variable (the result you are looking at) and 1 or more independent variables, whose effect on the dependent variable you are measuring. For example: you could choose Team Offensive Efficiency as your dependent variable (the result you are looking at) and Team Defensive Efficiency as your independent variable. You would be measuring the relationship between those 2 variables; more precisely, you would be measuring what you would expect Team Offensive Efficiency to be given a Team Defensive Efficiency.

R^2 measures how related the two variables are--if a team is good on O, does that change our expectation of what the team's D probably is? If R^2 is 1.0, they are completely related; if you know 1 you know the other for sure. If it is 0, they are unrelated.

First of all, thank you for your answer, it was clear and simple and I really appreciate it.

The reason I asked about regression is that I am currently working on a simple prediction model based on Pace, Off Eff, and Def Eff. I found two existing threads:

1.)http://www.apbr.org/metrics/viewtopic.php?f=2&t=8114
2.)http://www.apbr.org/metrics/viewtopic.php?f=2&t=8216

In the first one there was a primer about how Ken Pomeroy calculates his “expected scoring margin” and basically it is the backbone of my project. I created an Excel spreadsheet and played around with the numbers (adjustments for HCA and fatigue) but sometimes it is still under the game total by 10 or even 20 points.

Then I found the second one where the conversation is all about how Mr. Pomeroy calculates his adjusted offensive and defensive efficiencies. I realized that the adjustments he makes may be the key to solve my problem. Unfortunately, I am yet to master SLR, WLS, and Iteration. Which one of these would be the easiest for a rookie?

Thanks,
Ivan

(BTW, DSMok1, I love your Tableu Public presentations, they really help me visualize the stats.)