Data: How to hack esports’ biggest issue
Martin Dachselt is the CEO & managing director of Bayes Esports. He has over 20 years of top-level management experience in tech startups, previously serving as CTO of Dojo Madness, Smartfrog and Delivery Hero, and as VP operations for Click&Buy. Esports is exciting for him because of the high level of accessibility, with many fans also players themselves. Martin, who describes himself as the “old dude” of the startup scene, serves with his (life) experience of Bayes Esports as a beacon in this stormy and ever changing industry.
As esports continues its rapid growth, stakeholders such as data specialist Bayes Esports are working behind the scenes to build out a professional infrastructure to support its advance, and help establish it as a betting sport. Key to this, Bayes chief executive Martin Dachselt says, is accurate and official data.
Data is a disorganised collection of facts, he explains, which need to be manipulated into structured information. Despite the top leagues in esports being organised when it comes to collecting and presenting this data, the mid-tier segment is still relatively chaotic – and this needs to change.
To help the industry meet this demand, Dachselt says operators need to examine the current situation. He says there are ten key challenges to improving the esports data ecosystem – each of which the supplier looks to address with its proprietary data platform.
Ten key challenges
At Bayes they have identified that the current situation has several complex nuances that are causing issues. These include the uneven playing field in esports, integration of games, lack of a central source, accessibility, lack of organisation, live data streams, limited data, unofficial data, exclusivity and monetisation.
The uneven playing field describes the distribution of exclusive details and third-party products. When it comes to integration, several data sources have to be incorporated from people, systems and formats. This, combined with the lack of central sourcing hinders product innovation and efficiency within the industry providers.
Accessibility and the lack of organisation is another issue, where the absence of a central structure creates a bad user experience. Some companies have tried to tackle this by installing dedicated apps, however the risk of unofficial data consequently increases.
As Dachselt explains, “data is only available from a handful of tournaments and has to be integrated from various different sources using various different formats”.
“Add to this the various different tournament organisers, tournament formats and different operational processes and creating even relatively simple data-based analyses became an unnecessarily complicated and time consuming process,” he adds.
With esports being in its relative infancy compared to other forms of betting, it is crucial for operators to recognise that there is no central organisation or structure for data. Regional and national formalities do not exist, which means operators have to police themselves to some extent. As Dachselt emphasises, professionalism and trust is therefore crucial for development.
Consequently, navigating this space has potential reputational risks. “Many people still have presumptions and prejudices towards esports and, as such, esports cannot afford to appear unprofessional, as it could drive people away from it again,” Dachselt says. “We need data to be professional and we need to be professional for esports to reach its potential.
“Data and high quality information are at the foundation of professionalism. Without data, all analyses, game theories and all betting odds would simply be guesswork. A player’s performance cannot be accurately evaluated if there is no data to compare it to. Win probabilities cannot be calculated if there is no information about other matches.”
Live data and its risks
Live data comes from a few sources currently, Dachselt continues: live streams; broadcasts; official data feeds; and scraping websites.
The majority of games are broadcast via online platforms such as Twitch and YouTube, and although new technology has improved accuracy and filtering systems, there are still challenges when balancing the interests of viewers, punters and tournament organisers.
“We can’t emphasise how important it is to understand broadcast delay and its implications: esports broadcasts are intentionally delayed from 45 seconds (for tournaments with players on-premise) to up to 10 minutes (for tournaments with players at home),” Dachselt says. “Even the display screens in tournament venues are delayed by up to 20 seconds.”
Dachselt says it is therefore important to reduce risk by using the fastest available data – official server data – as the basis for odds making. This does create an even longer gap between odds and broadcast, which he admits can cause confusion for punters.
Visualisation or fast video streams can be used to solve this problem, but can affect the integrity of a match at best, this potentially leads to “spoilers” for viewers – at worst it could lead to matches being manipulated for betting-related purposes.
But this does not make unofficial data a viable option. It may be cheap, but it is inaccurate and unreliable.
Dachselt says there are four key risks involved with using unofficial data streams: speed; availability; granularity; and legal issues.
“Unofficial data sources might seem to be the cheaper option when compared to official data, but tend to cause additional issues later down the line that cause them to be more costly in the long run,” he explains.
“From bettors being able to exploit the slow calculation of betting odds, to customers losing interest due to repeated mistakes. This could lead to inaccuracies and potential legal actions from data rights holders against the scraping of data. Ultimately, unofficial data causes more harm than good.”
Platform for success
Esports data needs to be delivered faster to its clients than the stream can broadcast the action. If it is not, betting odds cannot be calculated accurately and visualisations that provide additional information to the viewers and that enhance the viewing experience cannot be made while the game is still running.
Dachselt says Bayes have tackled this through basing their live data services on official live data only via a low-latency, data-agnostic distribution platform. “We ensure that our data platform Bayes Esports Data Exchange (BEDEX) is the most technologically advanced platform in the industry,” he adds.
“By making sure that the data we provide to our clients comes directly from the source and through a platform that is always one step ahead of our competitors, we can guarantee our clients will receive the most accurate and reliable data as fast as possible.”
Dachselt says that BEDEX is paramount to Bayes Esports’ integrity. “We have developed real-time extraction of virtually any data point available (e.g. position of players). The data is cleansed, validated, enriched and transformed into proprietary data.
“For example, we add calculated data points and synthetic events, which are not provided by the game or API of the rights holder, but which simplifies integration for the consumers. We shield the consumer from changes of the data structure on the ingestion side to avoid that a game patch will break the implementation on the consumer side.”
The future of esports needs to include standards and neutral marketplaces like other industries. With huge data sets, innovation around analytical tools is becoming a key business strategy for operators.
It is imperative that gaming companies keep ahead of this curve through self-regulating and dispersing the turbulent landscape of data collection. For esports to become an official and mainstream form of betting, mid-tier segments need to be able to compete. And this can only happen if leading operators confront the issues to do with data.