9 Questions with Matthias Kullowatz

The pursuit of wisdom in any walk of life quickly reveals that what you think you know is not nearly enough to get you to where you want to go. As I'm starting out in my football journey I challenged myself to reach out to those already working in various roles in football to answer a short list of questions. My goal wasn't to get answers but relevant perspectives on the game within the game.

Here is Matthias Kullowatz:

What is your first memory of football?

I'm going to cheat and share an early memory of football. I grew up in Oregon, USA, where, in the 90s, "soccer" was growing, but it still was not very popular at the competitive level. Every kid played when they were 6, but by the time I was 12 or 13 there were only a handful of us left playing on competitive teams in my town. You couldn't just go to the park and find a pickup game to join...that is, unless it was basketball you were looking for. 

When I was 9 we lived in Ankara, the capital of Turkey. We lived across the street from a public school, where just about every night kids of all ages would go to play pick up games. I had never seen kids so competitive about soccer. Even the kids who drew the short straw to play keeper would go out there and dive to make saves. And we were playing on rough asphalt. The goals were marked by trash cans, and the ball was rarely pumped up. 

It didn't matter. This is what Turkish kids lived for: to make it through the school day so they could go play soccer at the school at night. I returned to the US the next year a considerably better soccer player, and ended up playing competitively through high school. There's a good chance I never would have played that long, or been as interested in soccer as I am now, if not for that year spent in Turkey.

What attracted you to data and scouting? What’s more intriguing now, refining your processes and acumen or 'discovering' players?

Definitely the thing that got me into data was baseball. There have always been so much data available from baseball, that it's the perfect training ground to learn statistical and data analytics skills. I read Moneyball (Michael Lewis) in undergrad, and decided to get a master's degree in statistics purely so I could better analyze baseball and become a GM. That didn't fully work out for me...yet. 

I would have to say I'm pretty equally intrigued by learning better tools to study sports data, and by developing methodologies to discover players. You could say that our recent creation of Goals Added (g+) combined both. I used xGBoost, a machine learning algorithm I've recently started using more, and we developed something that could very well help identify some of the next great players.

What is the biggest misconception/cliche regarding recruitment you’ve found in this space?

I mean, one of the most obvious ones is using goals and assists to evaluate players. There are so many lurking variables related to team strength and finishing noise that make those metrics vastly overrated. But more generally, I think there is still this rift between the analysts--nerds, if you will--and the front office. 

Many believe that you can't measure football with stats, and most of the rest believe that people who try to use their eyes to measure football are idiots. I think the less popular truth is that these things need to be combined. At ASA we have a wide range of bright minds. Some have watched a lot of football and a lot of players, and they have trained their eyes to catch important things, and others couldn't name half of Barcelona's current starting lineup (maybe that's just me?), but can code and build analytics tools. 

Most are somewhere in between, with a blend of those skills. When we come together and discuss a project like g+, there is so much synergy. We go through so much productive iteration, asking questions, checking the data, checking our intuition, and doing it all over again. Goals Added would not be as good a model if just I had created it myself. From measuring pass receiving value to using a multiple-possession time horizon to evaluate actions, the model would have come up woefully short of expectations without a mix of technical analytics skills and well trained football eyes.

If you could start over what skill would you build on first?

For me, I'm not sure I would do it much differently. My strengths are data analysis and model building. If I were restarting today, in 2020, I would immediately try to get more comfortable with cloud computing. At the very least, using cloud resources to tune models quickly or clean large datasets. I am getting old and I missed these things the first time around--maybe because they weren't there, or because they weren't very popular.

What is more important domain knowledge or curiosity?

Between my day job and working with the ASA team, I can definitely say both are important. But to me, curiosity is more important. Domain knowledge can be picked up, especially by curious people. And being curious leads you through explorations that develop not only more domain knowledge, but greater skills in data analysis and problem solving.

What is more useful in analytics - breadth of skills or depth of talent? Why?

I think this really depends on your team. If you are a lone wolf, you need a breadth of skills, and even when you have a team, a breadth of skills still helps late at night when you live on the west coast your team is asleep. But the most efficient work gets done with a diverse team of people that have expertise in different areas. It becomes an impediment to progress when you're putting the full burden on yourself to learn everything. 

At ASA, I don't necessarily have to be an expert in football tactics. As someone who played competitively in the distant past, I can still understand when someone tells me how a team is playing and where they're attacking, but I'm no expert. 

I can use the expertise of any one of our ASA tactics experts--John Muller, Kieran Doyle, Cheuk Hei Ho, Eliot McKinley, the Twitter personalities known only as JMooreQuakes and TiotalFootball, to name a few--who suggest where to look in the data for meaningful signal, and then I can code it up into something meaningful. I don't have to be able to set up a database and link it up to our R Shiny application. Because Tyler Richardett. 

You might start asking, what does Matthias even do? And that's a good question. A diverse team has the potential to be so efficient and do so much, but that doesn't necessarily require each individual to be diverse.

What three (3) football icons would you want to have a meal with? Why?

Mia Hamm, for one. She was, and maybe still is, the most recognizable face in American soccer. Not only is she one of the greatest football/soccer players of all time, I remember hearing about when she hit a few long field goals (American football) with her right foot, then hopped over to the left side and hit a few more. Field goals are hard! She's a boss.

I'm not sure about the other two. I'd probably go back home and visit some of my coaches growing up. It would be fun to catch up.

What advice would you give to someone wanting to get into this space?

There are some obvious ones, like learn to code, or learn to use video software really well. Basically learn to do something technical that many people can't do. Then to build on one of my other answers, make some friends in the community. There's a football analytics community out there with a lot of nice, helpful people. Eventually you might be able to slot your unique skills alongside theirs, and then you have 1+1 = 3.

What is your favorite quote or saying?

I have always been partial to: "Stop worrying about the world ending today. It's already tomorrow in Australia." Charles Shulz reminded us to relax every once in a while.