Computers are here to stay. The days of Pong are long gone, and many program languages today are far more sophisticated. They use Python modules that integrate powerful algorithms and historical data to predict impactful events. Some current programs even try to decipher the crazy stock market.
On top of that, there are programs that play sophisticated games like poker. And it seems that poker-playing programs are some of the most sophisticated inventions within the last decade.
Programmers use AI techniques that combine the power of terabyte storage and machine learning to make the best possible decisions. How could a human being compete with that? Hopefully, this will give you some insight.
Below, you’ll find seven facts about poker-playing computer programs.
1 – Early Stages of Interactive Computers
Competitive computer interactive gaming began as far back as the 1940s with the Turing chess challenges. These computers used simple algorithms that would only be as good as they were programmed. They had no learning capabilities.
These were only used by supercomputers that took up an entire room. And they failed at beating anyone with real skills. Success would not happen until the 2000s, when the gaming community drove the advancements of interactive computers.
2 – Gaming Begins to Hit the Public
Advances began at a consumer level in the ‘70s. Then, the ‘80s came about, and the gaming industry took a different turn with interactive games. Stories were developed, and football games with x’s and o’s ran rapid across Netscape and the World Wide Web.
During the ‘90s, computers began responding, and graphics allowed a person to become engaged into a battle with the machine. One of my favorites was Twisted Metal, where you were not only playing against opponents but also interactively against the machine through the internet.
In the 2000s, the advances of the ‘90s graphics combined with common internet access developed stories and competitiveness across the internet. Most of this required cable into public Wi-Fi, and security began to increase the capabilities of transferring funds. Banking became possible at the consumer level, which opened many new possibilities for gambling.
By 2006, 80% of US banks offered online banking, and online casino-generated revenues increased exponentially over this same period. These opportunities led to the first no-limit poker bot competitions. In 2004, the International Conference on Cognitive Modelling Tournament made online poker-playing computers a real possibility.
3 – Poker From the Human View
The game of poker is human in nature, and it goes beyond mere probabilities. The human factor takes stage with a sheer will to win. This game can make anyone do the most outlandish bets.
“I have a 45 suited, and I always win with that, so I’m hanging on at least until the river to get that big payoff that no one sees coming.” These are thoughts that seem great in the heat of the action, but do they always work out?
No… But what happens if it does? A person’s mindset for this type of behavior can make anyone using specific probabilities stark-raving mad. Comments like, “I can’t believe you did that,” and “You’re so lucky!” vibrate across the room.
Computers couldn’t really understand the human factor in poker by using logical decision-making schemes and, until recently, they were pretty easy to beat.
4 – AI Begins to Take the Stage
The advent of machine learning has revolutionized our society, and poker has not been left out.
Data scientists—a combination of programmer and mathematician—talk about Bayes Theorem, Support Vector Machines, Random Forest, and Neural Networks forming working brains within the computer.
It’s no longer the basic algorithm of probabilities used by the computer which blindly bets into poker extinction with no vested interest.
These new processes use probabilities that are finely integrated into general and specific historical data. Yes, it can gather my historical data as I play, which the bot puts into a data frame. This data frame now knows the types of bets I made, what cards were on the board when I made that bet, what I had at stake, if it was a raise, if it was a fold, and what cards and play the bot has made relative to my actions.
As you can see, bots are voracious when it comes to data collection. And just for good measure, a bot will associate my poker name with the data it has collected. Oh, and let’s not forget the date and times this single observation falls into the data frame as well.
5 – Computers Begin to Win
Contrary to what most people think, the computers today do not count every possible hand based on probabilities while learning my actions based off their data frame.
I have now been classified into a specific class of player within its algorithm. With a memory capacity vastly superior to mine, how can I stand a chance against the bot?
This data contains data from the equivalent of 5,000 hands of poker games, which it uses to create classifications. These classifications are based on detailed data from more hands than I can imagine being part of, and if I was able to imagine it, it may well have been forgotten.
6 – Computer Stamina
The thought you may have is, “I can figure it out if I watch and see what the computer does, then I know what it will do next time.”
Unfortunately, the computer brain is doing the same thing, except its brain never gets tired. It has a steady stream of electricity which never fails to fuel its calculations.
You and I, on the other hand, need to eat, sleep, and limit the substances that negatively affect our brains or judgment. I have to face the fact that, in the long run, I will wear out, get tired, and do something foolish. Then the bot wins.
So, if I choose to play a poker bot, how can I possibly win? There are ways to win against the bots, and more is better.
When I play with more people in the game, the computer has a challenging time understanding everyone’s strategy. The algorithms cannot understand humans. (Most of us are challenged in the same way.) So, the trick is to never play a bot one on one.
7 – Spotting the Poker Bot
When playing in a group, it’s important to pay attention to spot the bot, especially in the beginning. They usually play consistently from a time perspective with no real delays between folds, calls, and raises.
They’re not big chatters and rarely, if ever, comment. So, I will try to initiate a chat with an easily understood question to each player. No answer means they are on my watch list, and I look for other chats from the group to see if they communicate. It’s a little game of spot the bot.
As I mentioned earlier, they never get tired. When playing a specific site, I need to pay attention to players who are ALWAYS online. If they are there for hours on end, or perhaps days on end, you’ve probably spotted the bot.
Finally, is this user playing an unreasonable number of tables at once? I’m not sure about you, but for me, playing 10 tables at a time is a drag, if not next to impossible.
I’ve enjoyed this jaunt down the computer gaming progression. And I realize that it’s much more fun for me to play real money poker against humans than against a well-oiled machine. I hate to admit it, but I simply do not have the stamina nor the memory to enjoy playing against a bot.
The four things that help me spot the bot are:
- Play timing
- Chat messaging
- Extensive duration of playing time
- Multiple games at once (not just the typical few)
Gaming sites are not fans of bots and have an obligation to keep the game fun. Spotting and banning bots is just one of the things they do to justify getting a rake.
When I suspect there’s a bot, it’s my responsibility to report it to the poker site. If a site I frequent does not seem to take any action, I move my business elsewhere.
It’s far more fun and a lot more fair to play poker against people, not bots.