In 1995, Alexey Khobot designed his first software for FONBET, which included the use of bots to assist in oddsmaking. The operator looks at why this artificial intelligence has been key, but points out that the human touch still remains king.
The most important part of the bookmaking process is the formulation of odds. It’s no secret that bookmakers rely on the accurate calculation of odds to turn a profit. As a result, more and more bookmakers are relying on AI to assist them in their efforts.
Alexey Khobot designed the very first software used by the FONBET betting company. Now he is fully engaged in the development of computer programs for use in the sphere of bookmaking. That being said, can we expect AI to replace traditional oddsmakers in the near future?
How do bookmakers come up with odds?
A person without experience and an aptitude for mathematics will never be able to devise the exact odds for a sporting event. To do this accurately, you need to process all of the available information concerning each of the teams or participants taking part in the event. Statistics, form, results from previous meetings, injuries, suspensions – all of this must be taken into account. Then there’s also public opinion, and the bookmaker’s margin to consider. But what if some kind of force majeure suddenly takes place mid-event? How is all of this data handled?
Many companies don’t see the point in bothering to work out the odds they offer, they buy ready-made b2b solutions instead. In essence, they relay the data found on their sites from an external supplier. However, large companies prefer to employ their own professional bookmakers.
Robots in the betting industry. Why are they needed and will they replace people?
“Nowadays, we use robots for live betting at Fonbet, but matches are still overseen in semi-manual mode. What I mean by that is, each match is monitored by a responsible bookmaker, he revises the computer’s odds and is ready to react in case of force majeure,” said the company’s Marketing Director Alina Yakirevich.
What kind of force majeure situations are we talking about? VAR is an easy example. The referee may initially signal a goal, but then a few minutes later could change his decision after reviewing footage of the incident. VAR can influence the awarding of penalties and players being sent off.
But what if there is a fight among the fans in the stands or something like that? A bot doesn’t take this type of situation into consideration, but it could certainly influence the outcome of an event. In this case, bookmakers are able to temporarily suspend the acceptance of bets until the situation is clarified.
Alexey Khobot used to be a bookmaker himself. These days, he develops IT solutions for the betting industry, for example, bots that formulate odds and keeps them automatically updated as the match takes place. He created his first software for FONBET in 1995.
Khobot developed bots for the betting industry. These are programs that analyse incoming statistics and devise live odds for hundreds of options at the same time for one match. His first bot was for working out volleyball odds, and since last year he has been developing bots for esports.
Robots in esports – the spirit of the times
Esports is gaining popularity, and during a pandemic with real sports being suspended, the rate at which its audience is growing is impressive. Alexey Khobot wrote his first esports odds-bot for CS:GO. This game has an older audience compared to Dota2. In addition, this game is easier for a beginner to pick up.
Writing a bot for calculating esports odds is no easy task. There is an even larger amount of events to consider in CS than in classic sports. Each of them can affect the game. All of this must be calculated, with the probability of each event set. In addition, a team’s in-game economy is a very important factor. Which weapons a team buys, how it uses its budget, etc.. If one of the teams gains an advantage at the start of the game, it doesn’t at all guarantee that it will hold onto it.
Another additional difficulty to contend with is that the game is constantly updated, new maps are added, and this means the bot must also be frequently updated. It’s akin to the rules of football being changed every two months.
It took about a month’s work to create the bot. During this time, Alexey studied the game and watched dozens of Twitch streams. Testing the bot took another few weeks. Khobot invited a esports bookmaker to assist him. The specialist used Khobot’s program, feeding back with even the smallest of inaccuracies discovered in the odds it calculated. The CS:GO bot allows a single bookmaker to rule over several events simultaneously, while reducing the risk of human error
Alexey Khobot’s next goal is to devise a bot for live betting on Dota2. New meta is released monthly for Dota, making this a much more complicated task. Each update brings a huge change in the powers and abilities of the characters, and there are more than 100 of them in Dota. In addition to that, there are a huge amount of events that can take place in Dota, with many of them unique to the 10 individual characters on the game map.”
There is also an issue with data and analytics. As far as Dota2 is concerned, is difficult to find a reliable provider of data and reliable data is precisely what any bot needs to function correctly. Alexey firmly believes that these problems will be solved as esports continues to develop. Once this is done, he will be able to write his program for Dota2 and other esports. There is no way that AI can completely replace people in the betting industry. At least not any time soon.
Machines will not completely replace people
However things turn out, it’s unlikely that we shall see a person completely replaced by AI in the near future. Bookmaking concerns bets between players and traders. There is often a large amount of data that computers still struggle to analyse and digitise, especially with the biggest events. This could include rumours surrounding the fitness of players, or an athlete’s mindset coming into the event.
This is precisely the sort of data which AI is incapable of processing on its own.