In the first part of our conversation with ‘carlosthedwarf?’ we go through his journey with data analysis in sport, NRL Supercoach and views on the wider sport tech market in Australia. Check him out on Twitter for more great content & Part Two of the interview if you enjoy the article below!
Tell us about yourself
I’m based in Sydney; my background is in data and market analysis which shouldn’t surprise anyone. I started out working with retail sales tracking data across several industries and now work in the public sector doing the same thing.
How did you get into analysing sport data
I’ve always been a huge sports fan, mainly basketball, cricket and rugby league, especially the data side of it, with fantasy sports being the gateway. I think I participated in my first NBA auction draft when I was 13 and was in several NBA/NBL email fantasy leagues. Trying not to date myself too badly, I spent some time in the late 90s/early 2000s writing about the National Basketball League for a few websites and Pro Basketball Today if anyone remembers that. I used to look at box scores for hours on end, I loved it.
Once I started with data analysis as a career, I’d use small sets of data from different sporting leagues as training data for learning new skills and techniques since I knew what to expect from the outputs. Then I discovered NRL Supercoach, which as a tragic Eels fan was an outlet for suffering through decades of disappointment. Through that I ended up finding NRL Supercoach Talk, and submitted a few articles which got me started down the path of really digging into NRL statistic to find an edge.
Which sports do you analyse and how do you structure your analysis
Mostly rugby league these days, since I’m contributing to NRL Supercoach Talk but I’ll occasionally look at some NBA trends if something interests me. I try to look at data differently to everyone else or answer a question that would interest me that I don’t see anyone else talking about. Especially from the supercoach/fantasy sports side of things in rugby league, almost all the analysis is on a player or sometimes team level.
When dealing with retail tracking, I’d look at the size of the market, is it growing/shrinking and then drill down to a product type. Then maybe manufacturer level or brand level with the bottom level being individual products. Most analysis is based on individual products, the players. Is Andrew Fifita a better fantasy option than James Fisher-Harris? I just applied that retail tracking thinking to how I look at this data. What’s the average score for a Supercoach player? It’s about 49 for starting players depending on the season. What’s the average score for a front rower? About 52. And then you go from there. It’s just perspective, I look at it top down instead of bottom up.
if we want to advance the use and evolution of analytics here there needs to be more public data and more public analysis being done. Give it to the masses
What are your thoughts on the Australian sport tech environment and how we fit into the global landscape
The main thing that is holding us back is public availability of data. The technology and platforms are there but the data isn’t, at least to the general public. Anyone who follows NBA twitter knows the incredible variety and depth of analysis that has been undertaken by a wide array of people, and part of that has been from the NBA providing a lot of their tracking and advanced data to the public. That availability sped up the understanding and knowledge of the game among fans and media and helped usher in the current style of play.
We’re still many years behind here (if not a whole decade), and there’s always the issue with leagues and media partners here heavily controlling everything. We know there’s significantly more data that the clubs have access to, and the NRL runs a Datajam every year with outside people coming in to analyse their tracking data. I don’t expect them to release everything, and more data isn’t always a positive unless it’s useful. But if we want to advance the use and evolution of analytics here there needs to be more public data and more public analysis being done. Give it to the masses.
Everything public is volume based right now, which is only ever going to show that more is “better”, which guarantees you’ll never get anything more than superficial analysis. It’s one area that the punting and daily fantasy sports seem to be leading.
What advice do you have for the couch fan who wants to start playing around with some data
The only advice I can give is to just get started, especially learning how to automate the basic stuff to free up your time. Not only will you expand your knowledge of whatever you’re analysing, the other positive is that it’s a skill that employers value right now.
Which area or specific problem within sport would you love to look into further
Putting together the data for Tackle % in rugby league has made me really focus on trying to analyse quality defenders in rugby league. I’d love to get my hands on some positional data and take a deeper look at players who end up out of position consistently and how often a player inside/outside ends up covering for them. Teams are probably already doing that internally, but it’s something that really interests me currently.