Expert-Level Artificial Intelligence in Heads-Up No-Limit Poker
Libratus received the Marvin Minsky Medal for Outstanding Achievements in AI. Pluribus was on the cover of Science Magazine and was a runner-up for Science's Breakthrough of the Year for 2019. I was also named one of MIT Tech Review's 35 Innovators Under 35. Our goal was to replicate Libratus from a 2017 article published in Science titled Superhuman AI for heads-up no-limit poker: Libratus beats top professionals, and supplementary materials. Instead of building a poker bot for a full-sized HUNL Poker game, we scaled down to a Leduc game with 105 chips in the stack and 10-chip big and small blinds. Poker Genius is not only a game against AI opponents. It is a huge poker training complex which includes such tools as: Poker Hand Evaluator, Showdown Calculator, detailed Player Stats, Hand History Databases and Importer. You will feel the power of modern technologies and see how your poker skill is improving. Also in 2017, all top-ranked poker players were bested by software named Libratus from Tuomas Sandholm at CMU. The software adjusted its strategies during the tournament. And its algorithms for strategy and negotiation are game-independent, meaning they’re not just about poker, but a range of adversarial problems.
Links
Twitch | YouTube | Twitter
Downloads & Videos | Media Contact
Downloads & Videos | Media Contact
DeepStack bridges the gap between AI techniques for games of perfect information—like checkers, chess and Go—with ones for imperfect information games–like poker–to reason while it plays using “intuition” honed through deep learning to reassess its strategy with each decision.
With a study completed in December 2016 and published in Science in March 2017, DeepStack became the first AI capable of beating professional poker players at heads-up no-limit Texas hold'em poker.
DeepStack computes a strategy based on the current state of the game for only the remainder of the hand, not maintaining one for the full game, which leads to lower overall exploitability.
DeepStack avoids reasoning about the full remaining game by substituting computation beyond a certain depth with a fast-approximate estimate. Automatically trained with deep learning, DeepStack's “intuition” gives a gut feeling of the value of holding any cards in any situation.
DeepStack considers a reduced number of actions, allowing it to play at conventional human speeds. The system re-solves games in under five seconds using a simple gaming laptop with an Nvidia GPU.
The first computer program to outplay human professionals at heads-up no-limit Hold'em poker
In a study completed December 2016 and involving 44,000 hands of poker, DeepStack defeated 11 professional poker players with only one outside the margin of statistical significance. Over all games played, DeepStack won 49 big blinds/100 (always folding would only lose 75 bb/100), over four standard deviations from zero, making it the first computer program to beat professional poker players in heads-up no-limit Texas hold'em poker.
Games are serious business
Don’t let the name fool you, “games” of imperfect information provide a general mathematical model that describes how decision-makers interact. AI research has a long history of using parlour games to study these models, but attention has been focused primarily on perfect information games, like checkers, chess or go. Poker is the quintessential game of imperfect information, where you and your opponent hold information that each other doesn't have (your cards).
Until now, competitive AI approaches in imperfect information games have typically reasoned about the entire game, producing a complete strategy prior to play. However, to make this approach feasible in heads-up no-limit Texas hold’em—a game with vastly more unique situations than there are atoms in the universe—a simplified abstraction of the game is often needed.
A fundamentally different approach
DeepStack is the first theoretically sound application of heuristic search methods—which have been famously successful in games like checkers, chess, and Go—to imperfect information games.
At the heart of DeepStack is continual re-solving, a sound local strategy computation that only considers situations as they arise during play. This lets DeepStack avoid computing a complete strategy in advance, skirting the need for explicit abstraction.
During re-solving, DeepStack doesn’t need to reason about the entire remainder of the game because it substitutes computation beyond a certain depth with a fast approximate estimate, DeepStack’s 'intuition' – a gut feeling of the value of holding any possible private cards in any possible poker situation.
Finally, DeepStack’s intuition, much like human intuition, needs to be trained. We train it with deep learning using examples generated from random poker situations.
DeepStack is theoretically sound, produces strategies substantially more difficult to exploit than abstraction-based techniques and defeats professional poker players at heads-up no-limit poker with statistical significance.
Download
Paper & Supplements
Hand Histories
Members (Front-back)
Michael Bowling, Dustin Morrill, Nolan Bard, Trevor Davis, Kevin Waugh, Michael Johanson, Viliam Lisý, Martin Schmid, Matej Moravčík, Neil Burch
low-variance Evaluation
The performance of DeepStack and its opponents was evaluated using AIVAT, a provably unbiased low-variance technique based on carefully constructed control variates. Thanks to this technique, which gives an unbiased performance estimate with 85% reduction in standard deviation, we can show statistical significance in matches with as few as 3,000 games.
Abstraction-based Approaches
Despite using ideas from abstraction, DeepStack is fundamentally different from abstraction-based approaches, which compute and store a strategy prior to play. While DeepStack restricts the number of actions in its lookahead trees, it has no need for explicit abstraction as each re-solve starts from the actual public state, meaning DeepStack always perfectly understands the current situation.
Professional Matches
We evaluated DeepStack by playing it against a pool of professional poker players recruited by the International Federation of Poker. 44,852 games were played by 33 players from 17 countries. Eleven players completed the requested 3,000 games with DeepStack beating all but one by a statistically-significant margin. Over all games played, DeepStack outperformed players by over four standard deviations from zero.
Heuristic Search
At a conceptual level, DeepStack’s continual re-solving, “intuitive” local search and sparse lookahead trees describe heuristic search, which is responsible for many AI successes in perfect information games. Until DeepStack, no theoretically sound application of heuristic search was known in imperfect information games.
','resolveObject':','resolvedBy':'manual','resolved':true}'>
','resolvedBy':'manual','resolved':true}'>
','resolveObject':','resolvedBy':'manual','resolved':true}'>
What is Warbot
Warbot is Openholdem-based, customizable universal poker bot, which uses screen scraping method for its game state engine, and external profiles (formulas, algorithms) for its action engine (Autoplayer). The bot automatically detects poker table, when it appears on the screen. Then it connects and starts playing, according to loaded profile.
Along with primary actions (folding, calling, raising, etc..), it can also perform sitin, sitout, close tables, handle pop-ups, set custom randomizable delays and much more..
Features
- Plays No-limit and fixed-limit cash games, MTTs and SNGs (including Speed poker, DONs, Heads-up, Spin’n’goes, etc..)
- Based on the most advanced poker-botting framework that currently exists
- Supports 20+ largest poker rooms, the list gets wider constantly
- Plays solid position-aware TAG-style poker out of the box, packed with strong tested profiles for cash, MTT and SNG games. More than 50 extra profiles in our store to choose from
- Allows to edit existing profiles and create your own
- Adjusts in real-time to counter your opponents style of play
- Allows multi-tabling and running several casinos on the same pc
Results
- Warbot steadily earns money in micro-limit cash games with the average win-rate of about 5 big blinds/100 hands
- Even more impressive monthly profits in rakeback payments
- More than $20,000 signup bonuses to clear
- Dominates SNG tables: regular or turbo
- High money finishes and hitting final tables in MTT games completely unassisted
- Thousands of dollars to be made in tournaments with 60%-80% ROI (Return On Investments)
Room | Flags | Network | Tablemaps and table size | Security level |
---|---|---|---|---|
Stars | 3/6/9 seats default size | High | ||
888 Poker | 888 | 2/6/8/9/10/Blast default size | High | |
888 Poker.es | ||||
Party Poker | Party | 2/3/6/8/9/10 seats default size | High | |
Party Poker.fr | ||||
Bwin.com | 3/6/9/10 seats mini size | |||
Bwin.fr | 2/3/6/8/9/10 seats default size | High | ||
Bwin.es | ||||
Bwin.it | ||||
Winamax.fr | Winamax | 3/5/6/8/9 seats mini size | High | |
iPoker (ALL rooms with new interface) | iPoker (new interface) | 2/3/6/8/9 seats default size | Low | |
Betmost Poker | iPoker (old interface) | All kinds of seats mini size | Low | |
Bodog Bodog88.com | Bodog | 3/6/9 seats default size | Low | |
Bovada (Ignition) | 3/6/9 seats default size | Low | ||
CoinPoker | Private | 6 seats default size | Low | |
SealsWithClubs | Private | 6/9 seats default size | Low | |
PPPoker | Private | 6/9 seats real-money tables | Medium | |
Global Poker | Global | 3/6/9 seats default size | Medium | |
Carbon Poker | MergeGaming | 2/6/9/10 seats mini size | Low | |
Betonline | Chico | 6/9/10 seats default size | Medium | |
Sportsbetting | ||||
TigerGaming | ||||
PokerMatch PokerDOM | 6/9 seats default size | Medium | ||
Stan James | Microgaming | all seats mini size | Low | |
RedStars | ||||
RedKings | Medium | |||
Brasil Live Poker | Chico | 6/9 seats default size | Low | |
Poker de las Americas |
Security Level
High – use “Stealth with Bring” (see Manual)
Medium – use at least regular “Stealth”
Low – safe, no need for any protection
Signup Bonuses
Poker rooms attract new players by offering them an additional amount of bonus money for signing up and making a first deposit.
When you make a first deposit, the poker room usually adds a separate bonus account with your bonus money in it. As you play, this bonus money is released step by step into your main account.
When you make a first deposit, the poker room usually adds a separate bonus account with your bonus money in it. As you play, this bonus money is released step by step into your main account.
The bonus amount is usually a percentage of the deposit (e.g. 100% of your deposit) and most times, there is an upper limit (e.g. 100% up to $500)
Rakeback
Rake is a fee that a poker room takes from each pot as profit. The amount of rake taken generally ranges from $0.05 to $3 per pot, depending on the limit that you play and the size of the pot. Rakeback is a loyalty program for online poker players. With rakeback you can get a refund of the rake payments that poker rooms have already charged from you.
Libratus Poker Github Game
Rakeback is necessary to show a worthwhile long-term profit for most cash games. Please sign up for your poker accounts through a reputable rakeback provider. If you don’t have a bonus or rakeback going for you, we recommend switching to SNG’s or Tournaments
Wanted to thank you guys for this amazing stuff. You bot really rocks! I’m learning so much just by watching how it plays..
This thing is too addictive ? I could never imagine before, that computer program can play like this.. Awesome. And by the way, Alex, thanks a lot for great and fast support.
Thanks for everything, you really stand behind your product and I will always support you in any way I can.
Libratus Poker Github Bot
Purchase Warbot
Libratus Poker Github App
Purchase Warbot full version, with advanced profile for all major game types, and use it without any restrictions