"Women lie, men lie
Women lie, men lie
Women lie, men lie
Numbers don't lie" - Yo Gotti, featuring Lil Wayne
A funny thing happened on Friday. I submitted my "daily notes" blog to my editor, as usual. Then, a few minutes later, he started texting me about something I had written. Here's what I had written, and the series of texts that ensued:
Text 1: "SRS = Simple Rating System?"
Text 2: "When you have a sec can you send me a sentence or two to explain what SRS is?"
Text 3: "Is there an average or high/low to give these numbers context?"
A couple of things happened there. First, I was describing a couple of bad teams, and wanted a quick way to quantify just how bad they were so that you, the reader, could easily get my meaning. However, my editor was pointing out that the method that I used, SRS (read about it here), is not universally considered "standard" basketball language -- instead, it's an example of a dreaded "analytic."
A lot of people don't really know what a given analytic term might mean, so using terms like SRS isn't an everyday thing.
But it should be.
Ultimately, the stigma and mystery behind "analytics" should be stomped out, because "analytics" are really just ways to quickly, easily and numerically characterize what's going on in the game. That's it.
Analytics are a way to increase the toolbox of ways that we have to understand and describe what is happening on the court. They can be used on a high level by GMs looking to improve their teams. They can be used by analysts looking to describe the quality of play that has already occurred or to predict what might happen in future seasons. They can be used in fantasy sports team building, in designing video game ratings, or in setting a gambling line.
In other words, analytics are a tool. And they are a useful tool because they can be used to learn, improve the quality of our understanding, improve the caliber of the product on the court, and ultimately increase our enjoyment of the game. And the more that terms currently considered "analytics" are just brought into everyday use, the better it is for everyone -- including you.
This is one of the goals that I have for my weekly Monday column. I want to shine a light on what we call analytics, so that everyone reading can see that there's no hocus pocus to it. There's no egg-headery necessary. You don't have to have a Ph.D. to use them. And if you do start seeing and using analytics in your everyday sports life, the next thing you know, your own understanding and enjoyment of the game will get better.
So, to get started, today let's talk about the general types of analytics there are. For the NBA, I'll break it into three main categories:
1. Box scores
Each of those categories contains a lot of worthwhile information, so for the rest of this article, let's stick to a few box score analytics. See, one thing that many might not consider is that any number that you use to describe a sport is technically a type of analytic.
If you say that LeBron scored 31 last night, or that the Rockets have won seven straight games by at least 14 points, or that the 2016 Warriors won 73 games, you are using basic analytics already. And literally everyone who watches sports uses at least this level of analytics. These types of numbers are the basis for every sports conversation -- ever. And a lot of those basic analytics are found in the box score.
So-called "advanced" analytics, then, are attempts by people to utilize information found in the box scores in a more useful way. If the Warriors play the Thunder, and both Stephen Curry and Russell Westbrook score 31 points, but Curry does it on 17 shots while Westbrook does it on 33 shots, the positive value of their points in that game is different. In this example, Curry was much more efficient in getting his points, using a lot fewer possessions and, thus, allowed his 31 points to be more valuable to the Dubs that night than Westbrook's were to the Thunder.
We might characterize the difference in their scoring efficiency with a tool called true shooting percentage (TS%), which accounts for how many shots a player takes and makes, how many of those shots were 3-pointers, and how many free throws that a player has taken and made as well. The league average true shooting percentage last season was 55.2 TS%.
In our example, saying Curry had a 71 TS% while Westbrook had a 52 TS% would quickly and easily get across the point that Curry was the much more efficient scorer on this particular day, without any other descriptive words being needed.
But wait, the game is about more than scoring efficiency. Let's take our example further. In this hypothetical game, Westbrook also has 17 assists, 13 rebounds, six steals and four turnovers while Curry had seven assists, five rebounds, two steals and three turnovers. Using only this information, found in the box scores, we can say that Westbrook had a much higher usage than Curry, using the usage rate stat that adds up all of a player's points, assists, turnovers and missed shots to characterize how many possessions that player is using and, thus, how much weight he's carrying for his team.
So, if two dudes in a barbershop are talking about that game, that battle between the past two NBA MVPs, they might say:
Dude 1: "Yo, Steph and Westbrook were nice last night. Both dropped 31 in a close game."
Dude 2 (who didn't watch the game): "Who was better?"
Dude 1: "I'ma say Steph, cuz his shot was on. His true shooting percentage was up over 70, it was crazy. Westbrook had to work way harder for his shot. I will say, though, Westy was carrying way more load for the Thunder. Triple-double, like usual. Son had to do everything for his squad -- again -- just like last season when he turned in the highest usage percent ever."
That would be a real, legit, barbershop-level conversation. And the analytics would both fit in the flow of the conversation and make the conversation better, because Dude 2 would be able to get a clearer idea of what happened in the game without having to guess based on his boy's eye test.
Like the hip-hop mantra points out: Women may lie, men may lie, but numbers don't lie.