Sports Analytics 101 is a series of blog posts outlining the core concepts behind sports analytics in non-technical terms. You can find all available installments in the series here.
Sports analytics revolves around metrics: the numbers that go into a piece of analysis, and the numbers that come out on the other side. In fact, sports analytics can be broadly defined as the process of creating and analyzing metrics in relation to sports. If you don’t know how to think about different metrics and what they’re meant to represent, you’ll have a hard time doing sports analytics.
Later in this series, I’ll cover some common types of metrics and illustrate their usage through examples, but before we get to that, I’m going to introduce a framework for how to think about an individual metric.
Think of this framework as the mental “paperwork” to fill out whenever you use a metric. By ensuring you’ve filled out the paperwork, you can feel that you have at least a basic grasp of what the metric is and, just as importantly, what it isn’t.
In filling out the framework, we first name the metric, state what the metric is intended to quantify, and answer a set of questions about the metric. Over the next several posts, I’ll explore each of these questions individually, but first, let’s start at the top.
At the top of the framework, we lay out the two most important pieces of information: what the metric is and what we're using it to quantify. It’s vital to consider both of these pieces of information and not simply focus on what the metric is, because a single metric can be used to quantify different things. It’s impossible to reasonably judge a metric without establishing what the metric is being used to quantify. A metric might be great at quantifying one thing and useless at quantifying another.
Let’s consider a non-sports metric: the number of hours that Brian spends reading in a year (we’ll call this metric “Reading Hours”). Reading Hours may or may not be a useful or complete metric, depending on what we’re using it to quantify.
If we use Reading Hours to quantify how much time Brian spends reading, then obviously it’s a fantastic metric. On the other hand, using Reading Hours to quantify how well Brian understands American history may be a fool’s errand without more context. Is Brian reading about American history? If so, how well does he digest and remember what he reads? We need a lot of additional context if we’re to use Reading Hours as a metric to quantify Brian’s understanding of American history and even then, it may not be all that helpful.
When we’re using the framework, we’re not applying the framework to a metric, but rather to the combination of a metric and whatever it is we’re using the metric to quantify. To continue our example, we don’t simply judge Reading Hours as a metric on its own, we either judge Reading Hours as used to quantify how many hours Brian spends reading or Reading Hours as used to quantify Brian’s understanding of American history.
Once we’ve established what we intend to quantify with a metric, we can begin filling out the rest of the metric framework, asking questions like: Is the metric, when used in this way, a fact or a proxy? Is the metric adjusted for opportunity? The next several posts will delve into these questions and their place in the metric framework.