AI Poker Bots Everything You Need to Know
Table of content:
- Decoding the working of poker bots
- What makes bots tick?
- The New Bot Generation
- Will the game of poker be destroyed?
With a dose of luck thrown in, poker is a game of skill. That's if you're nice enough. It is a game with plenty of upswings and downswings for most casual online players and Return On Investment (ROI) that will do well with a few extra Christmas presents.
But just as monitoring software has given players the edge over their competitors, so bots are threatening a notch or three to step things up.
Monitoring software for online players was once the ultimate weapon. The bots have taken it to a whole new stage. And it's getting smarter with apps.
Better usability of software and data has contributed to improved statistical techniques for anything from sports betting to the stock market. Unsurprisingly, that has now spread to a few No-Limit Hold'em hands.
But can the psychology of a Will Kassouf needle or a massive dose of Phil Ivey violence ever be replicated by software? Will you really get rich playing with a bot online? Let's look at poker bots in more detail.
Poker bots are software pieces (usually) used on online poker sites by players who normally can't beat the game. Bots are computers that attempt to beat real-life human players using mathematics and player expertise.
Bots run alongside the app you're running in the background and as standalone programs. They mainly monitor the hands that have been played and make observations that can not be seen by the human eye.
It's like a HUD, but it also plays for you, actually. It's pretty quick to run a bot, as long as the poker site you're playing on allows them. They're all downloadable programs, and they come with simple instructions.
However, that has not prevented certain sites from cracking down on them. In the 2016 Turbo Championship of Online Poker, PokerStars investigated a supposed high-profile case of bots. Eventually, the player was found to be human.
It later emerged that bots were used by a group of Russian players to win over $1.5 million on the web for PLO cash games. The largest poker room in the world later made modifications to remove all third-party apps from its tables.
Initially, bots were very easy, and even if you didn't know you were up against a bot, it wouldn't be a challenge for an average player to beat them. As time goes by and technology progresses, however, things are beginning to change.
Modern-day bots are capable of using so-called solvers for stats and complex calculations. In today's games, solvers play an enormous role as they represent the players' final learning tool.
A poker bot can "remember" virtually endless numbers of pre-calculated scenarios, meaning that it can generate a correct answer in several different circumstances.
Bots are becoming a serious danger with this kind of basic information.
As Hold'em gets closer to being a "solved" game, these bots gain ground over human players more and more.
And while some sites do a fair bit to try and tackle the problem, it is impossible to quickly detect all bot play.
A human player may also make use of Bot-like apps. They may have the program running alongside the poker client (or even on a completely different machine, making it practically impossible to detect any kind of detection), giving them all the information.
So, even though you're not technically up against a bot, you're playing against that bot effectively. It might be a human being who clicks the buttons, but the bot tells them exactly what to do.
The human player is just an auto-clicker in this case. Still, using automated tools, it is much harder to track and detect.
Researchers are developing new bots that are better than individuals effectively. In specific situations, they not only play optimal strategies, but they can also be hard and, most importantly, bluff just like humans.
The Game Theory Optimum (GTO) strategy has advanced in recent years. This is an approach to the game, regardless of other players' actions, seeks the best plays possible in every scenario.
If you can play GTO, you don't have to worry about what other players are doing – you will profit.
Of course, if you know a little bit about Hold'em or PLO, you're aware that these are very complex games, so a human can't remember what the GTO game is for every possible situation.
Not because of a bot, though.
Bots don't have memory issues and don't get lost. They can get a perfect decision for every scenario so long as they run on a computer with sufficient storage space and processing power.
And it's a dangerous thing.
One reassuring fact is that GTO is still not perfect-Hold'em has not yet been completely solved-meaning that bots do not have access to a perfect and unbeatable strategy.
That said, things are changing quickly, and there are better and better solutions out there for poker-playing apps.
The game of poker has been seen as an outstanding AI testing ground.
Poker, particularly Hold'em, has many difficult quantifying aspects, unlike chess, which used to appreciate this word. If you feel like it, you can fold the best hand, and you can jump all in to win a tiny cup.
Of course, these are not always your best choices, but if you so choose, you are free to return to them.
It is this random variable that appeals to the specialists. At any given point in time, the number of options a player has is practically infinite. Therefore it has consequences beyond the green feeling to build an AI that can handle the mess and come out on top.
Such an AI may be used for complex simulations and for conditions with extremely unpredictable results to come up with solutions. It is obvious why such software in the military, medicine, strategic planning, and much more may be quite useful.
But the aim is easy for those running poker bots: to make money from unsuspecting matches.
The new bots have proved to be very successful against humans. While the best of the best can still hold their own, it seems like it is becoming harder for humans to keep up with the technology.
Whether these bots would absolutely ruin the game is the biggest worry many players have. Of course, it's hard to make assumptions on what could happen in the future, but it's fair to assume that we're not there yet.
Some poker bots concentrate on tables for cash games. MTT players do not have to deal with many bots, and the reasoning for this could be that the AI behind them is not built to cope with poker tournaments' ever-changing nature.
Players in cash games have it tougher.
It's no secret that, even at very respectable stakes in some poker rooms, bots are sticking around. But it is primarily against these poker robots that the operators take a stand. Many of them are likely to do their best to defend real players.
Finally, the poker game will change, and in the future, it will almost definitely change.