Is Poker Really Dead in 2025? The Truth Revealed

14 Apr 2025

Intermediate

This material is for medium-skilled players

From discussing the difference between strong and weak GTO solutions to dissecting the challenges of RTA-based cheating, this trendy topic gives you the tools to make informed decisions about your own poker journey. We also touch on which formats might be safer or more lucrative in the coming years, and what trends could shape the next chapter of the game. Whether you're a recreational player or someone taking poker seriously, this episode aims to replace fear with clarity. The game isn’t dead — but it is different. And understanding those differences could be the key to finding long-term success.

Good morning, and welcome back to another episode of the Red Chip Poker Podcast. I’m James Sweeney — you might know me as Splitsuit. In this episode, we’re diving into a question that’s been floating around the poker world for years: is poker dead? You’ve heard the arguments: GTO solvers taking over, RTA users cheating, insane rake structures, unbeatable bots… some say all of this has killed the game. But is that really true? Or is there still room to profit, evolve, and stay one step ahead?

In this article, we’ll separate myth from reality. We’ll look at the evolution of poker AI — from Cepheus solving Limit Texas Hold’em, to the ongoing challenge of solving No-Limit Hold’em. We’ll also talk about whether today’s GTO tools are actually making players stronger or just more confused.

By the end of this episode, you’ll have a better grasp of the real threats poker faces, how to adjust your own poker strategy, and why the game is far from dead. With that said, here’s Coach Weasel to break it all down — enjoy!

If too many players start following GTO solutions, that raises a serious concern: «Can we still make money playing poker?». To get to the core of it, this question often comes down to the potential existence of RTA users — that’s real-time assistance. We’re talking about either bots running GTO solvers in real time, or human players who are clicking through solver outputs while they play.

So the real question becomes: «Does the presence of RTA cheaters make it impossible to generate consistent income at the tables?». The concept of real-time assistance isn’t new. In fact, it predates the commercial release of solvers. Even if you go all the way back to the pre-Black Friday era — before 2011 — there were already players claiming poker was dead. 

«Everyone’s too good to beat», they said. And after Black Friday, when the environment arguably got tougher, even more people jumped on the «poker is dead» bandwagon. This narrative has followed poker through every era, right into the modern game. And along that journey, there’ve been a few milestones that really solidified this fear for a lot of players — that maybe we’ve reached the end of profitability in online poker.

One major event? The 2015 announcement that a poker AI named Cepheus had solved poker — specifically, heads-up Limit Hold’em. For players already worried about the game’s future, this felt like the final nail in the coffin. We all knew this day would come. Poker is solved.

So with the announcement of Cepheus, the message seemed clear: «There’s no way to make money playing poker going forward». Here’s the key point — Cepheus can’t be beaten over a large enough sample. Even if you spent a massive amount of time studying the GTO solution for heads-up Limit Hold’em, the best-case scenario is that you break even against it. And in a raked environment? You’d both lose over time — even playing perfect strategy.

So no, you cannot profit playing against Cepheus in a raked heads-up Limit Hold’em game. That sounds pretty bleak. Naturally, after hearing news like this, it might feel like online poker is doomed. But it’s crucial to keep these kinds of announcements in perspective.

Let’s talk about what it actually means to «solve» poker — and how relevant the Cepheus announcement really is. First, solving poker isn’t black and white. There are different levels of solving, and that’s something we’ll explore in a second.

Did Cepheus mean the end of online poker as we know it? Not really — and here’s why: Cepheus was only built to play heads-up Limit Hold’em. Think about it — when was the last time you played heads-up Limit Hold’em online? Heck, when was the last time you played any form of Limit Hold’em, six-max or full-ring? Your answer is probably «a long time ago» or maybe «never».

Sure, back in the day, Limit Hold’em was more common — but in today’s poker world, it’s all about No-Limit and Pot Limit games. Fixed Limit has mostly fallen out of favor. Even if there are still some hardcore Limit Hold’em grinders, most don’t play heads-up. They’re in six-max or full-ring games.

Technically, what Cepheus achieved is called a weak solve. Now, this gets a bit deeper into game theory, but it’s important. There’s actually a hierarchy when it comes to solving games — and it’s not always discussed clearly. You can find this breakdown on the Wikipedia page for «solved games».

There are three levels of solutions listed — and while it’s unclear whether these are accepted as mathematical fact or more of a thesis (they come from a PhD dissertation), they’re still useful:

• Ultra-weak solution.
• Weak solution.
• Strong solution.

A strong solution means the game has been fully solved — where an algorithm can make the perfect move from any position.

Here’s the key point about a strong solution: it works from any game state — even if mistakes have already been made by one or both players. That means even if the current situation is theoretically impossible in perfect play, a strong solution still knows the optimal path forward from that exact point.

Now, step down one level, and we get a weak solution. A weak solution knows all the correct plays, but only under the assumption that both players have played GTO up to that point. In other words, it can respond perfectly from any point that naturally arises in a GTO game tree.

But if one of the players deviates from GTO — say, they make a random or exploitable play — then the weak solution might not know the best counter. Only a strong solution can do that. And that’s what makes Cepheus a weak solve for heads-up Limit Hold’em. It knows how to play perfect poker — but only from GTO-correct positions.

Sure, it’ll still probably win (because it’s playing solid GTO), but even Cepheus doesn’t know everything about every possible hand state — just those that follow perfect play.

Now, let’s talk about the third and lowest tier: the ultra-weak solution. This kind of solution doesn’t describe the full strategy. It just proves whether the first player wins, loses, or draws with perfect play from both sides — starting from the initial position.

Think of it like this: imagine a heads-up No-Limit Hold’em game, and we prove that the player with position has a long-run advantage. That’s an ultra-weak solution. It tells you the eventual outcome, but not how to get there.

And there’s a spectrum within ultra-weak solves. Some are very basic (like the example above), while others can get quite complex — detailing large chunks of optimal strategy — but still fall short of qualifying as a full weak solution.

Now let’s bring it back to No Limit Hold’em — where most of today’s poker is played. We’re all familiar with solvers. Any serious player can access one. You can load a spot, punch in ranges, and within seconds get a GTO approximation of what to do.

So, the obvious question is: where does solver output fall on this spectrum — strong, weak, or ultra-weak? Well, solvers are not full solves. Not even close. What they give us is best classified as an ultra-weak solution.

When we compare solver output to the solution Cepheus provided for heads-up Limit Hold'em in 2015, the difference is staggering. Cepheus qualifies as a weak solution because it takes into account every possible action as part of a perfect GTO tree, including all different bet and raise sizes.

This is relatively straightforward in Limit Hold'em because there's only one bet or raise size allowed per situation. So, from that perspective, we could describe the heads-up limit hold'em tree as small. But «small» is a relative term — this is still a huge tree. Consider all the possible flops, turns, and rivers in Limit Hold'em.

To give you some perspective: Cepheus, a supercomputer with 48 CPUs, took 68 days to solve heads-up Limit Hold'em. Now, contrast that with No-Limit Hold'em. The game tree here is orders of magnitude larger because in No-Limit Hold'em, each situation could involve a wide range of bet sizes — anything from a check to an all-in, and every possible denomination in between. This exponentially increases the complexity of the game tree. Even with the most powerful supercomputers, it's unclear whether we’ll be able to reach even a weak solution in our lifetime for No-Limit Hold'em.

Once you've spent the computational power to find a weak solution for 100 big blinds, you’d have to repeat the process for 150 BB, 200 BB, and so on. The scale of the problem is immense. So, how do solvers even work? How do we get close to GTO output?

The answer is that solvers heavily simplify the game tree. If you’ve worked with solvers, you’ll know the process starts by setting parameters for what bet sizes are allowed at each decision node. For example, on the flop, you might tell the solver it can either bet half-pot or the full pot. The game tree is simplified to the point where there are only a few possible actions at each decision node — sometimes just one or two bet sizes. At times, the solver may not even have the option to bet; you might force it to check rather than make a donk bet, for instance.

Now, let’s return to 6-max No-Limit Hold'em. It’s unlikely we’ll ever see a strong solution or even a weak solution for this in our lifetime. But that doesn’t mean solvers are useless. On the contrary, even an ultra-weak solution — what we essentially get from solvers — is still incredibly powerful.

While the term «ultra-weak solution» might sound unimpressive, it's actually quite potent: most players won’t be able to beat even this basic level of GTO output. What’s more, GTO solutions are widely available today. Pre-solved outputs, solver software, and training tools are all within reach for anyone.

This raises an important question: even without a complete solution to No-Limit Hold'em, has online poker profitability already died if cheaters can access these solutions in real-time? If the average player can’t beat an ultra-weak solve, does that mean poker is effectively dead, even though we don’t have the full solution?

These are some of the considerations we need to take into account when thinking about real-time assistance cheaters and the future of online poker. First off, using real-time assistance (RTA) is against the terms of service of most poker sites. While this may seem like an obvious point, it's not always straightforward to just start cheating. Poker sites have dedicated security teams, and unless you go about it in a highly sophisticated manner, you’re likely to have your account flagged or frozen. I’m not even sure the average person knows how to stay under the radar of these security teams.

When this happens, they’ll start investigating and banning players suspected of cheating. So, for what it's worth, there are safeguards in place, but there are other aspects to consider beyond just the security team.

There are two main types of RTA: automated and manual. Automated RTA refers to running a bot with solver output, while manual RTA involves using solver outputs to make decisions manually during play. Manual RTA sounds simple — just look at the solver’s suggestion and act accordingly. However, it’s much more difficult and time-consuming than it might seem. Even with perfect knowledge of the game, poker is complex. Implementing a mixed strategy consistently requires a high level of skill. 

Plus, constantly pulling up solver models, consulting grids, and manually selecting actions for every hand takes a lot of time. This approach doesn’t generate high win rates, and even playing a weak GTO solution can result in low profits, especially after accounting for rake. Most high winners in poker are exploitative players, not GTO purists. So even if you're following a solver’s weak solution, you’re likely looking at modest win rates at best, especially when playing low-volume, mentally intense sessions.

Who would be willing to grind that out for 12 hours a day, running the risk of being banned from the site for not flying under the radar? It’s a method that requires a very specific type of individual, and it’s not something most players would pursue. Ironically, it might be easier to get a real job than to invest that much time into manual RTA.

Automated RTA might seem like a better option, but consider this: Do you have the technical skills to code an RTA bot? Can you set up a system to integrate your solver with a poker client, automating decisions in real-time? It’s not something you can just buy off the shelf, and if someone has created such a system, they likely want to keep it for themselves, using it to maximize their own profits.

They don’t need a bot to succeed—they could likely outperform most players at high stakes simply by using their own skills. While it might be tough to beat someone who’s using an RTA bot, such bots do a poor job of evading security teams, and they come with significant barriers to entry — technical know-how and hardware that most players don’t have. This means only a small percentage of players could even attempt to cheat in this way.

That said, it's not necessarily a problem if there are better players in the game. We all know we can't be the best player in the world, and there will always be players better than us. This doesn’t prevent us from making money in poker, though, because a large part of poker success comes down to game selection. If you suspect a player is using RTA, it's best not to play against them frequently. Instead, focus on opponents you can outplay.

As long as you stick to this strategy, the presence of a few cheaters doesn’t have to negatively impact your profits. You don’t need to be the best player in the room — you just need to be smarter about choosing your games.

Another thing to consider is that No Limit Hold'em isn’t the only poker variant online. While players might find ways to cheat at No Limit Hold'em using real-time assistance (RTA), this specifically applies to that game.

There are far fewer players likely to be cheating at Pot Limit Omaha (PLO), especially the five or six-card variants, or other more exotic poker types. In fact, there might not even be solvers available for these variants, making the likelihood of someone using an RTA bot in these games quite slim.

In a worst-case scenario, where everyone is using a strong solve for No Limit Hold'em, we could simply switch to a different form of poker — like PLO or another exotic format — that hasn’t been as heavily studied or solved.

It’s also worth noting that even if we were facing off against an RTA bot using an ultra-weak solution to poker, these types of solutions are technically beatable. They aren’t perfect, and if you know how to exploit them, it’s possible to generate a positive win rate against them. We’ll dive deeper into this in another episode of the Red Chip Poker podcast.

Something I often think about is whether the abundance of GTO (Game Theory Optimal) information available today is actually making players better. For a long time, there seemed to be a gradual improvement in the skill level of online players year after year. But with the surge in GTO resources, I’m no longer convinced it’s truly enhancing players’ abilities. While players now discuss more advanced, theoretical concepts than ever before, I'm not sure it’s translating into tougher competition. In fact, it seems that players may be more confused than ever.

We used to focus on more straightforward, exploitative strategies online, but now the game has become a complex mix of grids and frequencies. This is the first time in poker history where I’m genuinely uncertain whether the overall skill level of online players is improving — or slowly deteriorating.

To sum things up, is poker dead? Absolutely not — it's still alive and well. The decline of poker is far off, and if it were to happen, it would apply to specific variants. In fact, the bigger threat to poker isn't cheaters; it's large poker sites dominating the online poker market. For example, some sites charge unreasonably high rake on higher-limit games, which makes it hard for professionals to sustain long-term profits.

Certain poker rooms seem to treat poker more like a casino game such as roulette or blackjack, where they don't expect players to win. These sites don’t consider the long-term viability of poker for players, which could lead to problems for the industry. It’s essential for players to be selective about where they play, to avoid giving too much power to entities that don’t have poker’s best interests in mind.

In the worst-case scenario, if cheating or real-time assistance becomes widespread, the solution might simply be to shift to a different variant, such as Pot Limit Omaha. I hope this discussion gives you something to think about. Thanks for tuning in, good luck at the tables, and happy grinding!