Closing the loop – A5 Labs Head of Security, John Andress, explains how GeoComply is stopping poker cheats with WPT Global

And how does that apply to online poker?

JA:  The landscape of cheating has evolved over time and become much more sophisticated. 

We have a big interest in protecting the integrity of the game. It’s beneficial for us from a revenue perspective and the players want to feel safe. 

Around 2015, a lot of more publicly available study tools were released. PioSOLVER is the most common one.

It’s a reinforcement learning AI that creates a model of the game. Players use it to study, but players can also use it to create tools to tell them what to do in the game, which is the problem.

I used to play poker professionally and there has been a change in the online game. Player skill was more evenly distributed. Even among professionals, there were okay players, mid-to-high stakes guys, and the cream of the crop high-stakes players. Now the skill gap is more polarized, making it more difficult to make it as a professional.

So, the concern is that more and more people have access to the tools and will use them for the wrong reasons. We want to prevent that and build tools or at least offer, on our platforms, tools that can detect cheating or at least disincentivize the users from using them and make it expensive for them.

There was a cheating scandal with GG Poker, where some well-known high-stakes players were using tools like this not just for themselves but for some of the players they staked in the games.

Professional players have always been aware of the tools because they use them to study. But now, even the recreational community or the casual poker players, who really are the lifeblood of the business, are more aware of them.

These players want to know what operators are doing to protect them and ensure the game is safe.

So what are you doing to protect them?

JA: Back in the day, PokerStars was the gold standard. They had this sophisticated software to detect bots with mouse movements, but they were somewhat invasive on your computer.

For the most part, high-stakes poker players are okay with the software essentially knowing what they’re doing if it means that the games will be safe. But now that’s not good enough, and we’ve made a big bet on taking it a step further.

We want to make sure that players are not running some prohibited software, but we also want to understand where they’re playing from, maybe have some back information about them, and then the most sophisticated element of this is really understanding and being able to measure skill through how they play the game.

So that’s where we’ve made a big bet and hired a lot of industry experts in machine learning and data science, along with professional poker players. So we can not just look at the signal you would have if you were running a high-risk piece of software, like a virtual machine that is conducive to having this type of AI but also look at how you make decisions in the game.

And then, we compare your play to the population of pro players or the whole population of players to understand if we can see some signals through our prediction algorithms that show that what you’re doing is not exactly human.

Maybe you’re doing things a little too consistently or you’re just such an outlier from how people normally act, that we should investigate. 

So I would hope that some of the other operators have invested in this area, but I can’t say for sure because it’s been a big project.

So tell me about GeoComply closing the loop?

JA: For a while, we were really losing the cat-and-mouse battle with the so-called contextual detection.

When players changed their location, we didn’t know where they were playing from. We didn’t know what software they were using. We were always one step behind them. They might have been connecting from a particular country or region where we know these issues are prevalent, or they’re from sanctioned regions and we want to block that entire country.

This was outside of our expertise. They were very sophisticated in making it look like they were an organic user in a region that we service.

GeoComply makes sure we can verify players’ location and have that contextual identity, to have a complete understanding of their network behavior.

The SDK sits on the client side. It collects this information and ensures the customer isn’t manipulating our software. So we don’t think they’re doing one thing, and they’re actually doing another. If our analysts can trust the data, they can make fast decisions, and it acts as a preventative measure that makes the bad actor take a step back and forces them to invest more money and time to try to get around the technology.