Did you know a quick Google search for “poker training” turns up 741,000 results? From online videos to strategy articles, there isn’t a shortage of options. What is in short supply is variety. Most poker training is geared toward three of the four primary learning styles, those being visual, auditory, and read-write. What’s missing as far as online options go, at least until now, is kinesthetic learning.
Enter Poker Fighter, a hands-on poker training software designed by Guy Sela, programmer and developer of the software, and poker player Stas “Stasia42” Tishkevich, who also serves as Product Manager. Part simulator, part entertaining game, Poker Fighter utilises state-of-the-art technology to allow players the opportunity to play risk-free poker while getting immediate feedback and advice from the pros.
“You can play, practise and improve using this unique online poker trainer that simulates real cash games,” says their website, poker-fighter.com. “The game was designed to offer better access to the poker world for newbies and experienced players who want to maximise their winnings and improve their understanding of the game.”
How it Works
Poker Fighter, which is currently geared toward online $0.05-$0.10 and live $1-$2 games, is a guided simulator that puts users through hand-after-hand of different poker scenarios. Throughout, the user makes his or her usual decisions and is then offered feedback on each. However, by adding the ability to earn points, introducing characters, and allowing users to “Pick a Fight,” Poker Fighter has gamified poker training.
“The current ‘for dummies’ version teaches optimal poker versus loose and passive opponents that tend to call too much both pre and post flop,” says Tishkevich. “These types of players often limp preflop and rarely three or four-bet. Post flop they continue with their passiveness, rarely raise their draws, tend to call too much wanting to reach a showdown, etc. As a result, Poker Fighter teaches a somewhat conservative approach versus raises and re-raises and tends to bet wide for thin-value, while semi-bluffing with good draws. These are the sort of things players can expect to learn.”
Tishkevich continues: “We also have a ‘Pro’ version that teaches wider polarised ranges and teaches how to exploit loose and aggressive opponents by putting pressure on wide capped ranges. The poker brain itself can be easily adapted to various opponent types based on their known tendencies and ranges.”
Here’s a short video on how exactly Poker Fighter works:
Origins of Poker Fighter
In 2010, Sela, who at the time was serving in the army, had the idea of developing poker training software. Trying to master the game with limited access to poker training resources, Sela began to think outside the box, specifically about creating artificial intelligent software that could teach people how to play. The idea was simple -- develop A.I. software that adapts and offers feedback on your actual decisions.
Meanwhile, Tishkevich was crushing the mid-stakes cash games online thanks to his devotion to heads-up display (HUD) stats and mathematical analysis of ranges. In 2013, the duo met and decided to develop Poker Fighter, though, at the time the idea was to offer sophisticated pro-aimed software with advanced scenarios and ranges.
Two years later, an Alpha version of "Poker Fighter for Pros" launched, and while they received positive feedback from the pros, they neglected novice and recreational players, who complained that the scenarios and explanations were too sophisticated. Given recreational players are the biggest market share in poker, the duo adapted.
“By simplifying scenarios and explanations we managed to create a Beta Version of Poker Fighter ‘for dummies’ in early 2016, and got very positive feedback from our testers,” says Sela. “We invested some time and money on UX and UI and started a primary local marketing campaign. This month, we plan to start a new marketing campaign, targeting the rest of the world.”
Trials and Tribulations
Developing Poker Fighter wasn’t easy. One of the biggest challenges was taking the complex game of poker and analysing every possible scenario to translate it into language that computers understand.
“We had to program AI algorithms to give the software the skills of a pro poker player, allowing it to make correct decisions based on relative hand strength, board texture, possible holdings that the opponent might have, etc.,” says Sela.
While many believe that computers play poker better than humans, Sela and Tishkevich realised this wasn’t the case as most of the time programs fail to adjust to their opponent. With a lot of programs, there tends to be only one solution for each hand, which of course is not always the case. For instance, different strategies should be used against players that bluff too much compared to players that only bet with strong hands.
Among the poker decisions translated into code for Poker Fighter are:
- Relative hand strength pre-flop, post-flop, based on actions and positions.
- Ranges of hands based on actions and position.
- Board textures and their effect on ranges and strategy – Flop, Turn and River.
- Strategies for common scenarios – Isolation of limpers, applying pressure on scare cards, etc.
- *Adjustments to various opponent types (Tight/Loose, Aggressive/Passive, Folding/Non-Folding)
“Anything which can be thoroughly explained in words can be translated into code,” explains Sela. “No exceptions whatsoever. Some concepts are way harder than others to code, but that is because they are harder to explain to humans. Take an instructional video for $1,000NL in comparison to a video for $10NL. The explanations and analysis of a hand and a board in $1,000NL will be much more complex. Same goes for coding these different levels of thinking. It is feasible, but harder.”
The Technology Behind Poker Fighter
Unlike some other training simulators, the hands on Poker Fighter aren’t predetermined. By not pre-programming the hands and boards used in the game and instead allowing the AI brain to generate them as you play, Poker Fighter becomes an infinite game. As long as there are different hands and scenarios in poker, of which there are millions, there’s always more you can learn from Poker Fighter.
“We programmed its brain with enough poker skills to allow it to handle all the possible scenarios we wanted to teach,” says Sela. “The biggest advantage it has on other AI poker software is that it can actually explain why it recommends to do a certain action. It does not act as a ‘black box’ that just tells you ‘you should have raised.’ Instead, it will tell you exactly why you had to do it.”
Because they didn’t program it to solve every possible spot out there, Poker Fighter’s AI isn’t what online poker players call a bot. For those that don’t know, a bot is a computer program that plays without human intervention (it’s like playing chess against a computer).
“Poker Fighter’s teaching method divides the game into scenario groups -- preflop raising, continuation bets on wet boards, defence on the turn in two-bet pots, etc.,” Sela continues. “We optimised the poker brain to handle these scenario groups and didn’t code for groups that we don’t teach. We definitely have all the building blocks to transition Poker Fighter into a poker bot if we choose to do so.”
Poker Fighter’s “poker brain” veers away from traditional Game Theory Optimal (GTO) strategies to adaptive ones, which both Sela and Tishkevich feel is more profitable when playing against human competition.
“The biggest problem with GTO strategies is that, by definition, a GTO strategy is never an adaptive strategy,” Sela explains. “It never adapts itself to the opponent's game. The reason behind it is that if you would change your strategy according to what you have seen so far from your opponent, the opponent could potentially exploit you and change his own strategy just before you adapt to him.”
The problem is poker is a game between real players, meaning there is a human element. Poker Fighter accounts for this.
“Applying an adaptive strategy against humans, even though it's theoretically exploitable, will be the most profitable strategy in the majority of the poker tables,” Sela goes on to say. “We can let the game theorists continue researching for an unexploitable strategy in their ivory towers, while we apply our adaptive strategies in the casino.”
Plans for the Future
Poker Fighter is currently a web-game that is customised for both desktops and tablets. However, the technology was developed in such a way to make an easy transition to a mobile app. Indeed, an app is in the cards for 2017.
In addition, Poker Fighter intends to expand the types of games they offer beginning with nine-max cash tables, and then moving to heads-up No-Limit Hold’em. Other future possibilities include Poker Fighter simulators for multi-table tournament and even pot-limit Omaha.
“Our poker brain is very flexible, so it all comes down to writing down the various strategies, a challenge for the Product Manager rather than the coding expert,” said Tishkevich.
You can play Poker Fighter now (no download required) at poker-fighter.com.