Hello again, everyone; it’s been a while. Let me tell you a little story.
By the time I launched I//R last July, I had spent six months brainstorming about just what this site was going to be. (I know, it takes me a long time to make decisions.) In the end, I determined that it should accomplish four main goals:
- Publishing my own NFL research using advanced statistical techniques.
- Separating wheat from chaff in the online NFL analytics world.
- Uniting the academic and online NFL knowledge bases.
- Testing out all these non-NFL content ideas I have in my head
Soon thereafter, I devised a posting schedule that would serve all of these masters.
But then I changed my mind (again), and decided to go for a more fly-by-the-seat-of-my-pants approach. And then reliability analysis began receiving positive feedback, so I latched onto it like grim death. To mix metaphors, I was feeling my oats while going through growing pains.
With the onset of the NFL offseason, however, it’s time to return to Plan A, so here’s the posting schedule I’ll be abiding by on I//R for the foreseeable future:
- Modeling Monday
- Wisdom Wednesday
- Reliability Thursday
- Random Rounding
Let me briefly explain what each type of post is all about.
In these posts, I will use various types of advanced statistical techniques to analyze NFL data, all of which fall under the purview of structural equation modeling (SEM). I’ve already introduced multilevel modeling (MLM) here, and demonstrated it in this analysis of Pythagorean wins. I’ve also introduced latent growth modeling (LGM) here and demonstrated it in this analysis of quarterback age curves.
Both MLM and LGM have far-reaching applications in an NFL context, but the general SEM framework includes more than just MLM and LGM. Other SEM techniques I’ll be using are mixture modeling, which Boris Chen has used to create weekly player tiers in fantasy football, path analysis, which I used to test whether or not penalty differential is a mechanism of the NFL home-field advantage (HFA), multi-group invariance analysis, which I also used in my HFA study, confirmatory factor analysis (CFA), multiple indicators multiple causes models (MIMIC), and more.
As one can infer from the above paragraph, SEM is an incredibly flexible tool for modeling complex phenomena, and the anatomy of an NFL outcome is — if nothing else — complex.
These posts will take one of three forms:
- Critiquing metrics and studies one can find in the online NFL analytics universe; And I mean “critique” in the classical sense, not the colloquial sense, i.e., I’ll identify both negatives and positives, as well as issues that should be addressed going forward.
- Discussing NFL research that’s been published in academic journals. There’s a lot of good work that’s already been done, and someone needs to both shine a light on it and translate it into lay English. Stats in the Wild already does this to a certain extent, but they need allies. If not me, who? And if not now, when?
- Applying what I learned getting my Masters degree in sport psychology to what I see in the NFL.
These posts will continue the extensive line of research I’ve already done on I//R: “When does Stat X stabilize?” and “What’s the mandatory minimum for a player at Position A to qualify for the rankings of Stat X?”
One part of the randomness of these posts is that they’ll be untethered from any set schedule — at least aside from the fact that they won’t show up on Mondays, Wednesdays, or Thursdays. The other part is that they won’t be NFL-related. Rather, they’ll be my outlet for writing about other topics of personal interest; mostly through the lens of statistics, but sometimes not.
What topics, you might ask? Well, to preserve said randomness, I’m not going to spill all of the beans right now. That said, you can be sure that topics I’ve discussed on the Three Cone Drill Podcast (e.g., music, poker, and pro wrestling) will play a recurring role in Random Rounding.
DT : IR :: TL : DR
This will be my posting schedule on I//R going forward:
- Modeling Monday: Applying all manner of SEM techniques to NFL data.
- Wisdom Wednesday: Critiquing online research, illuminating academic research, and soapboxing sport psychology.
- Reliability Thursday: Determining stabilization points and mandatory minimums for NFL metrics.
- Random Rounding: Writing about whatever the hell non-NFL topic I desire.