- Plays Poker Automatically - card and stats input, calculations and mouse clicks fully automated.
- Artifical Intelligence -
- NEW Artificial Intelligence - Holdem Bot uses Neural Nets and Genetic Algorithms. The new version 2.0 has been trained on over 1,000,000 games of hand history.
- NEW Bluff detection - This is a totally new function which helps Holdem Bot to identify pure bluffs, check raises and slow playing players, and adjust the opponent modeler accordingly.
- Multiple tables - run the poker bot on unlimited number of tables
- Auto screen scrape - card recognition
- Pot odds - to call
- Implied odds - added to pot odds
- Chance to win - is mainly based on opponent modeling, player stats and simulations
- Opponent modeling - uses an artificial neural network for betting pattern recognition
- Player stats - memorizes the opponents' game style, used for implied odds and hand odds
- Simulation - games are randomly simulated to identify profitable decisions (Check-raise etc.)
- Setup area - to optimize the play on different tables
- Semi-Auto mode - auto folds bad hands in preflop. Works on both No-limit and Limit tables!
- Random click time (0-3 sec)
What is Semi-Auto mode good for?
When you're playing poker you often find the folding in preflop a very time consuming procedure. It feels like this because preflop doesn't require much thinking. Still preflop is the most important part of Texas Hold'em.
The Semi-Auto mode takes care of the boring folding procedure for you, and makes a ring signal when you've finally got a decent hand. Holdem Bot won't interrupt once it has made the signal, so then it's up to you to make the decisions. Still you can glance at Holdem Bot's conclusions during the whole game.
How does Holdem Bot come to a decision?
Holdem Bot calculates your chance to win and compares it to the pot odds. On the basis of this comparison it comes to a decision.
How does Holdem Bot calculate my chances to win?
When Holdem Bot (HB) has collected information about your cards, your position, your opponents' bets etc, the calculations may begin. The betting pattern of the first opponent is fed into the neural network, which is trained to assume the probability of an opponent's 101 subsequent hand strengths (0%-100%). HB performs this procedure for all opponents and pick out only the 1-3 strongest opponents. HB stores these values into a hand-strength probability-table.
Now, it's feasible to calculate the probability that an opponent has a certain pocket hand. For example, if we suppose that our opponent has JJ in the Flop, and calculate that he at this moment would have a handstrength of 30%, then we can, thanks to the hand-strength probability-table, find out the probability of our opponent having JJ. HB iterates these comparisions with all the remaining cards and generates a card probability-table for the 1-3 strongest opponents.
Finally HB has just one, but still very complex procedure left; to evaluate all possible card outcomes, for the strongest opponents and for the next betting round. All outcomes for an opponent are weighted proportionally with the card probability-table. The result is calculated with the formula; Win-Loss+Tie/2 divided by all outcomes. The result - an accurate value of your chances to win.