The game we all love needs some help. Recent cheating scandals have proved that more needs to be done to weed out and identify bad actors. And long-term growth can only come from attracting new players and convincing them to stick around. That will only happen if the games they’re playing are fun, fair, and beatable.
A5 Labs has a solution for all of the above – a novel blueprint for player profiling that extends far beyond game security, suggesting a potential revolution in everything from table selection to loyalty rewards. That involves profiling players not just on “risk” but on “value” too. The full details are available in its recently-published white paper titled Advanced Technologies for Secure, Fair, and Fun Online Poker. Here’s how it works.
How player profiling technology works
There are three key ideas for player profiling in the A5 paper. There is “contextual data” for very broad, very rough categorizations. There is GTO hand analysis for pinpoint precision. And finally, there is machine learning to drive efficiency and greater accuracy.
Contextual data is an assessment of your internet footprint as a poker player. Things like what kind of device you’re on, where your IP address is located, and whether you have other software interacting with the poker platform. This would also include broader information like your win rate and the most common stats like how often you enter a pot and how often you three-bet from the big blind.
Data like this allows the sites to do a quick risk assessment. Colluders in the same building are likely to use a VPN to disguise the fact that there are multiple logins from the same IP. So, using a VPN would raise your risk. A player with a VPIP of 30% from under the gun is unlikely to be using RTA.
This contextual information doesn’t tell the sites who the cheats are, but it quickly cuts away the kind of players who are just not likely to be bad actors.
With the pool rendered much smaller, it becomes easier to search for suspicious behavior. That is where GTO hand analysis comes in.
The future of poker profiling for security
Players tend to start learning GTO with the most common and simple spots and work towards solving and learning ever less likely spots. A5 Labs grades the spots they are interested in differently.
What the A5 team is looking for is low entropy spots. It wants the largest amount of information from the fewest number of hands.
There are plenty of spots, especially on the river, that even the most GTO-savvy human is likely to screw up. Cheats don’t have this problem. If a player is using RTA that contains solutions for these spots, they will outperform even the best player. To the A5 Labs algorithm, things like that stick out like a sore thumb.
“Contextual information is pretty limited, it is not really reliable but gives you some indication of a player’s risk,” says Thanh Tran, co-CEO of A5 Labs. “But with gameplay information, even with just two hundred hands, if you have a good GTO baseline it’s really hard not to make big costly mistakes unless you have RTA helping you. Especially when you get to the river. Even pros make mistakes here.”
The process of flagging risky players would be smoothed and automated through the increased use of machine learning using, as Tran says, “AI to build a much more robust system that can analyze all of this and output predictable and reliable results.”
The future of poker profiling for sustainability
As well as making games fair, they need to be sustainable as well.
Tran makes a good point here: “Amateurs can’t contribute sustainably to the poker industry because they don’t stand a chance,” he says. “The lifetime of these players is really limited. Most sites have about 5-10% of the players capturing 90% of the winnings.”
One solution is a sort of rating system. Take chess as an example, a game where most people would insta-quit if they faced skilled players from the off. You can log into chess.com, hit “play” and be paired with someone around the same skill level. In much the same way, A5 Labs pictures a future in which poker software sits you at a table with other players in your skill bracket.
This would help to keep beginners from dumping their cash too fast, allowing them to find their feet and step up as long-term members of the poker community.
Tran explains: “We think that technology has a big role to play. If you can identify skill level, identify places and issues that make the games not fair. That’s the broader context: trying to help a greater range of players actually win. We think in terms of a fairness quotient where players move closer to the break-even line, helping have a fairer distribution of wins and losses and to extend the lifetime of these players.”
According to A5 Labs, this isn’t something that’s designed to hurt pros at the expense of new players.
As Tran says, “We see operators devising smart incentive schemes to ensure their professional players continue to make a strong profit as long as they contribute to engagement, fun, action, and overall sustainability. We have been working on a loyalty scheme that rewards players that create that action.”
It’s a holistic solution where everyone plays a part for the good of the game.
What about rehabilitation?
The biggest fear with a system like this is false positives or a kind of irrevocability. However, Tran has reassurances on both fronts.
“I’d like to use the analogy with the credit rating system,” Tran says. “A stable and sustainable system has to take into account the short and long term. There are things you have done recently and things you have done historically, so you want the system to be weighted towards the most recent actions.”
In other words, time heals. Keep your nose clean for long enough and you can earn back a good reputation.
What the system really needs in order to allay fears of accidental permabans and the like is, according to Tran, “anonymity and a third party, an unbiased organization to checkmark proposed standards, again like the credit industry.”
And when we have that “credit rating” system in place, what then? Managing the huge amount of data and allowing operators to use it without compromising anonymity is a significant problem.
However, A5 has a solution for that too, NFT avatars, which we will talk about in the next entry in The Future of Online Poker series.
Join the debate. The A5 Labs White Paper, Advanced Technologies for Secure, Fair, and Fun Online Poker, sets out a holistic vision for a safer future for online poker. You can get more info about its long-term mission to make poker secure and sustainable at A5labs.co.