Where once ideas and services, including our very own gambling industry, were driven by physical, earth-bound entities, today—supercharged by the hard lessons of the Covid-19 pandemic—almost every facet of our lives is powered and enabled by online technologies trafficking in data.
Data has revolutionised gambling. Increasingly traditional gaming is transitioning to iGaming. So who better to explore and explain the employment and development opportunities offered by the data revolution than data industry leaders Sven Lund and Ashley Washington at Berlin-based Bayes Esports Solutions?
And with Sven, Bayes’ Vice President of Esports Data Services, and female trailblazer Ashley, Senior Product Owner (Betting), giving us their brilliant, cutting-edge, insights we were not disappointed.
As we move more and more into a digital age, Data Science has emerged as the key to success in our industry. How did Bayes Esports manage to create a hugely talented data team within such a competitive market?
ASHLEY: “I think that the key to finding the right data scientist talent in this industry was having a clear idea of what our team needed when we were hiring. Bayes Esports is a young company but we have a developing culture and, at its core, it’s really about a lot of people who love games and enjoy Esports. Finding a skilled data scientist isn’t the hard part. But finding a data scientist who is passionate about Esports–and what we are doing at Bayes– is another story altogether. This isn’t necessarily difficult but it’s sometimes tricky to find the right fit, the right individual, within such specific parameters. For example, someone who spends a lot of time both playing and watching competitions for the core titles may be more appealing than someone who just plays.”
SVEN: “Picking up with what Ashley said, I think that with Esports we have a compelling area within which we work. This content definitely attracts people who are interested in the field. Additionally, we have a healthy need for entry-level talent in Data Science. While there are many intensely complicated applications of Data Science out there, Esports applications benefit from tried and tested—and therefore more straightforward—methods of Data Science that those new to the field will have possibly already mastered.”
How can the iGaming industry better attract data scientist talent? Can we compete with seemingly more attractive B2C sectors?
S: “Are you kidding me? I think working with Esports data is inherently “cool”. For example, I modelled Dota 2 markets when I was still working as a data scientist and building a machine-learning model that predicts the map winner. Seeing such a model in action — one that hopefully also performs well — is a very rewarding experience. This is, therefore, a pretty easy case to make for new applicants if they are interested in Esports to begin with. If they are not, we don’t care.”
A: “I generally agree with this. I personally struggle sometimes to find adequately diverse options when I am hiring. I think there are many factors that play into this. While finding game-loving applicants isn’t often an issue, attracting women, for example, is always something that proves to be a bit difficult. I think this has a lot to do with the nature of the gaming industry to begin with. It always appears to favour male tech talent, so virtually any other industry will appear more attractive to women as a result, as long as it isn’t outwardly male-oriented. At this time, a clear effort has to be made on female-focused initiatives — or even deliberate one-on-one connections. It’s not enough to just put the listing out. We need to put more work into the issue. But I’m confident that improvements can already be seen across the board — and hopefully within Bayes as well.
“Regarding B2C being more attractive than a B2B application of Data Science, I have to say that B’s are often just really big C’s – in other words it can be really easy to immediately assume that a business application of Esports Data Science is going to be really boring. But our work moves at the pace of the industry and that cadence is often quite exciting depending on which competitions or new titles pop up along the way. The businesses that we serve often seek one consistent product but they also often develop a range of differing expectations and interests that guide future development in interesting ways. Additionally, the models Sven mentioned can potentially be tweaked to fit a variety of applications, including consumer-facing products.”
When hiring a data scientist, what is the best measure of their suitability for the role?
S: “Coding ability sits at the core of this challenge. Applicants can work on that from home so that they have the time to really think about their work and this largely eliminates stress as a factor. We review these challenges and this already acts as an important filter. The really interesting part is then to engage with the applicants in the interview about their work.
“And then we look at three key factors: 1) can the candidate explain the solution in clear terms; 2) is the candidate critical regarding their own work and 3) can the candidate think on their feet by responding to on-the-spot modifications to the original problem proposed by the interviewer.”
A: “Since we work for the same company, our processes are largely the same. The main difference for my own team’s hiring is that we also rely on a quick preliminary skills check that is mainly meant to determine whether or not the applicant possesses foundational skills in Data Science and Python. The coding challenge usually comes after they’ve passed that short quiz (mostly multiple choice) and have also spoken with me on the phone about the position and their expectations. I think having some face-time between these technical steps is really key to keeping an applicant interested and helping us to put a face to the name early on.”
How important are statistics-based questions when measuring aptitude?
S: “It depends. Let’s say, the applicant uses a certain machine-learning model and we ask why they used that particular model, or why it behaves in a certain way, then statistics might play a role. But you don’t need a PhD in statistics to excel at Bayes.”
A: “I’m inclined to agree with that. The candidate’s statistical knowledge is usually revealed in conversations about their challenge solution, but we don’t necessarily build those questions into the challenge. Depending on the version, stats questions might also be included in our preliminary check.”
What are some of the key lessons you’ve learned over the years when managing and developing Data Science talent?
S: “One of the most important lessons I’ve learned is that it is very important to strike the right balance between routine and challenging new topics in Data Science; although this isn’t something that’s possible to achieve at all times.”
A: “There is definitely a lot of truth in what Sven said. I think that it’s super easy to get swept up in all of the cool things you can do with Data Science, especially when you’re still learning. What I think challenges many newcomers sooner or later is that a lot of business uses for Data Science aren’t as inherently riveting as research work. I’ve spoken with many data scientists of all skill levels who express a great deal of relief and happiness when they really get to play around with a dataset or a problem outside of the constructs of business needs and deadlines.
“This isn’t to say that there are no intriguing real-life applications of Data Science. It’s just that often they’re not to be found in the standard business context. It can be a tough adjustment in perspective to make. Finding ways to introduce new opportunities for some research and ideation is something that I’ve found to be key to keeping otherwise repetitive work interesting. Thankfully, the gaming industry is, in many ways, up to the task.”
How is this infrastructure impacting Bayes’ growth?
S: “I’m not sure if I understand the question correctly, but if you mean to what extent does Data Science impact Bayes own growth, I would say it affects growth significantly. For example, if we discuss a new title or data input, the Data Science team is involved quite early on informing us about the feasibility of the undertaking in regard to predictions. Keep in mind that we either need to deliver probabilities based on the given data or that our customers need to create probabilities based on that data. Therefore, if Data Science would tell us that we can’t build models that make sound predictions, that would massively influence our decision about acquiring the data or not.”
A: “Absolutely. As a gamer myself, I’m always quite curious about the Data Science team’s outlook for the different titles. You really never know when your favourite game might be the next big thing in the market. It could be my chance to be the resident expert at any moment!”
Editor’s Note:
Well, remember you read it here first, as our industry moves, or is that elides, from the retail tangible to the iGaming ethereal.
The future of Data Science is assured, Sven and Ashley contest. Data empowers. Indeed, data is power.
But don’t forget, as they have eloquently reminded us, if you’re considering a career in Data Science most importantly you need genuine enthusiasm and real commitment, traditional values that remain forever necessary and key to work success.
And–despite the high-tech, cutting-edge lure–putative data scientists need also to embrace both the mundane and the marvellous.