Modern online sportsbooks are technological powerhouses in many ways. Lightning fast live betting updates, interactive live markets and seamless performance across devices and borders is no mean feat to build and maintain. In play betting makes up as much as 60% of betting handle at some major sportsbooks today, meaning operators cannot afford to ignore this part of the market. But what are some of the challenges in this kind of model?
To the user, a live bet is just a quick tap – and away they go. But the tech behind it goes much deeper. This article will take a fascinating look into the backend mechanics of live betting at an online sportsbook, from the basics of data architecture to how sportsbooks get their data delivered and how they deal with taking and settling thousands of bets at once at lightning fast speed.
Real Time Transfer and Delivery – How do Sportsbooks Get Their Data
Sports data providers of the modern day can push data updates as fast as every 200ms during fast sports matches. With so many player prop markets now available, the systems tracking player data are often just as fast as the points updates. Players often wear biometric and GPS tags for team performance analysis, and some of this data will be swung over to sportsbooks for prop bet purposes.
After the data leaves the live provider, the odds and results on the screen of the end user should update in a further 250ms or less. The entire process should be substantially less than a second in most cases.
Technologies like WebSockets connect a user to a server with an ongoing connection, rather than needing a reload to update with new information. There can be, during intense sports matches with lots of live betting, several hundred messages (or pings) a minute going back and forth between a user and the sportsbook server.
This amount of data requires careful balancing. Systems like caches and load balancers can stop a sportsbook’s server from being overwhelmed when there’s a lot of action going down.
The former being servers built for processing fast-evolving information on a temporary basis, and the latter being something like triage in a medical setting wherein servers are assigned data as it comes to keep any from getting too full.
With the rise of global markets, the 24/7 nature of live wagering significantly amplifies data challenges. For example, during a major fixture, the SpiderBets live betting section can attract well over 100,000 concurrent WebSocket connections, with similar spikes occurring multiple times per day across different regions.
Modern Sportsbooks’ Live Betting is a Great Case Study of Systems Data Architecture Basics
A modern sportsbook’s live betting system will be compartmentalized into various microservices. The alternative is a monolithic system, which handles everything in one place. This isn’t so suitable for live betting, because there are so many different services sending data back and forth that each process needs to be highly specified.
For example, consider a bet on the next goal in an NHL game. This involves the sportsbook’s app or website, the operator’s central server, the sports data partner and the real-time odds updating system and the bettor’s personal device, all within under a second or so. In many cases there will also be an API (or several) that acts as a translator between these different systems.
There may also, in some cases, even be geolocation tracking ensuring bettors haven’t moved out of an area not served by the app their using. This could also detect things like quick changing IPs, if a bettor moves between locations during a match – or something more suspicious like bots.
Having all these systems act independently also helps system uptime stay stable as if one things crashes it doesn’t take out everything. It also means things like user data can be stored in a separate system, reducing load on servers during peak time.
Fair Betting – Security and Fraud Prevention in Bet Settlements are Also Key
All of this goes on behind the scenes to ensure that the bettor gets a smooth, and error-free experience on their end. Lightning fast (literally) speeds are needed in live betting.
A bettor could place a bet on a goal just a half a second after the goal crossed the line, and if the system wasn’t fast enough they’d be paid out. People also try to automate this kind of thing, which can be detected through velocity checkers – software that monitors bet frequency and timing for suspicious patterns.
Once an event trigger has occurred, most sportsbooks will settle a bet in under 10 seconds or less. The final challenge is also long-term storage. Almost everything will need to be stored for a financial year, but some regulated sportsbooks in certain markets are required to store wager data for up to seven years.
This isn’t just for security or in case of dispute either – most modern sports bettors expect to be able to look back at their betting history and analyse their own patterns and stats and this is now a common feature.