About this project

My name is Dylan Wiwad I’m a quantitative behavioral scientist who spent most of my academic career as a social psychologist thinking about how to measure stuff you can’t directly observe—things like biases, beliefs, experiences, etc. Despite the fact that my substantive research area was in the psychology of inequality, I always loved and have done some original work in psychometrics. I found the problem of how we create reliable, valid, and meaningful metrics for unobservable psychological phenomena to be a really interesting puzzle.

I worked on a lot of questions around inequality and measurement from an academic angle. After moving into industry as a Quant Researcher at Slack, those same skills started creeping back in—but this time in more applied contexts like "How do we take the nebulous idea of 'AI Agent Quality' and build a solid construct definition and measurement model around it?"

And then, hockey crept back in.

I grew up in Edmonton, so naturally I was a hockey fan as a kid. But like a lot of people, I hit that point where I thought I was too cool for it. Plus, the Oilers’ “decade of darkness” didn’t exactly make it easy to stay invested. So for a long stretch, I just stopped paying attention. During grad school, sports crept back into my life through the road bike racing scene in Vancouver, and I started to re-enter the world of participating in, watching, and following competitive sports.

Then in 2021–2022, I got a text from my brother, a lifelong hockey player and diehard fan, saying the Oilers were making a legitimate playoff run for the first time in a long time. I got curious again. Watched the playoffs that year. We lost. Watched again the next year. Lost worse. Then I started dipping into regular season games. By 2024–2025, I was fully back watching every single game, with Moneypuck open constantly for live analytics.

At some point that season, I started thinking "hey, this Moneypuck data must come from somewhere..." A bit of digging later, I found the NHL API documentation and started exploring. I began pulling historical data, building dashboards to track shots, time on ice, and a simplified depth metric in real time. Mostly just for fun and to see what I could build.

Most hockey analytics out there are focused on raw observables or prediction; things like shot counts and shot locations, penalty minutes, expected goals, or win probabilities. That stuff is super interesting and useful, but it also misses an angle that is right up my research-psychologist alley: the aspects we can't observe, the latent variables. No one seemed to be trying to define or measure the squishier, more abstract concepts—the ones everyone talks about, but that aren't easy to see or define.

That’s when my psychometrics brain kicked in. I started noticing how often broadcasters and analysts would throw around terms like “depth,” “physicality,” “pace,” or “defensive responsibility” but no one ever really defined them. Everyone uses them, but I’ve never seen a clear, structured way of defining and measuring them.

So I started thinking: what if I could build metrics to quantify these kinds of ideas at the game or team level? What if I could turn nebulous hockey ideas into interpretable, useful data, and present them in a way that didn’t feel like a data dump? Most hockey analytics sites feel like they’re built for other data nerds. I wanted something clearer, more accessible, more narrative driven.

That’s how Hockey Decoded came to be. It’s my attempt to merge a career in measurement and data with a renewed interest in hockey. Part analytics, part storytelling, part exploration. The goal is simple: decode hockey. Make sense of the data in a way that is interesting to the people who just want to learn something they might not have known yesterday, but also deep, rigorous, and novel enough for the stats nerds.