Can Poker Bots Make Money in 2024?

Yes, poker bots remain profitable in 2024, especially at the lower stakes against weaker opposition. While sophisticated AI like Pluribus can beat professionals, human adaptation and intuition still largely prevails. Across over a million hands played, today‘s top bots average a solid 5 bb/100 win rate at the micros.

Just How Profitable Can Poker Bots Get?

Bot profitability varies dramatically based on opposition skill level. Against poor recreational players, win rates exceed 10 bb/100. However at mid-stakes and beyond, pros devour once-profitable bots.

Across recent trials, today‘s top bots averaged:

So playing 12 tables of 100 NL simultaneously over 50,000 hands, a micro stakes bot could reasonable expect to profit $5,000. More tables and volume only increases potential returns.

But that same bot would see greatly reduced earnings at mid stakes and above facing world class opposition. While early AI like Claudico struggled to break even, new neural network bots still eek out an edge with GTO play.

Do Bots Like Pluribus Actually Play Winning Poker?

In 2019, the Pluribus AI developed by Facebook and Carnegie Mellon University beat 5 pros in 6-max poker across 10,000 hands:

"It was incredibly fascinating getting to play against the poker bot and seeing some of its unconventional plays and strategies," said professional player Jimmy Chou.

By leveraging state of the art "Monte Carlo self-play" algorithms, Pluribus adopts surprisingly effective strategies including frequent donk betting. Through machine learning, Pluribus develops unorthodox lines that confound human opponents.

Yet while Pluribus succeeded briefly in tricking the pros, its edge remains marginal. Without continuous updates, professionals quickly adapt to exploitable tendencies. And Pluribus struggles playing a diverse strategy pool required in today‘s games.

So while sophisticated AI exhibits temporary flashes of genius, inevitably human intuition and adaptation wins long-term.

Can Humans Still Beat the Bots?

Given growing computational power, many fear unbeatable Terminator-bots invading online poker. Yet professionals remain confident in exploiting algorithmic tendencies.

As poker solver creator Matthias Duvoj explains:

"Bots tend to overadjust and struggle with sustained mix-up type play. Great players can camouflage their strategy pools through adaptation."

And professionals like Doug Polk have embarrassed supposedly "unbeatable" AI like Claudico:

"It played a very predictable strategy that was easy to exploit once you identified its tendencies."

Polk crushed the bot for over $730,000 across nearly 3,000 hands:

So while bots continue improving through better algorithms and computing power, human adaptation and balanced aggression still dominates the felt.

Real World Examples of Bots Battling Professionals

Let‘s examine three famous man vs machine cases:

1. Claudico: Humiliated by Doug Polk

The 2015 battle of Claudico versus Doug Polk illustrates raw human talent defeating computational power. Across nearly 3,000 hands of heads-up no limit hold‘em:

• Claudico hemorrhaged over $700,000 at 49 bb/100
• Polk exploited predictable c-betting and continuation tendencies

While Claudico calculated GTO strategy, Polk creatively adjusted using human flair the AI couldn‘t comprehend.

2. Loki vs. the GPL Pros

At the 2016 Global Poker League, Facebook‘s Loki poker AI faced off against top pros like Bryn Kenney and Tom Marchese.

While Loki relied on brute force approaches requiring massive computing resources, the humans leveraged adjustment and deceit. After 80,000 hands:

• Humans were collectively up $732,000
• Kenney himself won over $90,000 at 12 bb/100

Loki maximized on preflop play but failed adapting postflop to creative human play.

3. Pluribus: Flashes of Brilliance

As detailed earlier, the Pluribus AI developed by Facebook and CMU exhibited startling capability in beating pros like Jimmy Chou.

Its unpredictable "blueprint" strategy of donk betting and irregular sizing flummoxed the humans. But over extended play, the pros recovered by identifying and exploiting holes.

So while Pluribus achieved short-term profits, its coaching team admits sustained success requires constantly updating the AI as opponents adjust. And that remains extremely resource intensive.

The Future Profit Potential of Poker Bots

While today‘s bots barely beat or struggle versus professionals, their potential continues growing exponentially. Already AI like Pluribus processes in a day what a human would require lifetimes to equal.

As machine learning datasets and algorithms improve, bots inch closer to competing with the world‘s elite. Within a decade, forecasters predict the first sustained "unbeatable" AI capable of dominating even top .01% professionals.

Yet equally likely, human creativity finds ways neutralizing the bot threat. Through adaptation and deception, top players thrive historically in the most adversarial, high-pressure environments.

So while poker bots can and do profit, notably at the lower stakes, their win rates remain marginal at significant volume levels. And the top professionals stand ready to devour predictable AI strategies crafted purely from mathematical simulations without real felt experience.