Inside PokerSnowie's brain reveals the work of the Snowie AI Team. It explores first hand how the brain of PokerSnowie evolves and learns advanced strategic concepts, on its own.

PokerSnowie's ultimate aim is to produce the perfectly balanced game, find the ultimate un-exploitable equilibrium for all No Limit Hold'em configurations. Join us on this fascinating journey, which is just starting, into the future of poker.

The PokerSnowie Video Quiz series aims to answer the question: "what is the right play?" It is a set of poker coaching videos presented by French professional poker player Sharp. In each video, Sharp sets up an interesting hand in the "Scenarios" tool of PokerSnowie and explains how to analyse the situation and learn from PokerSnowie's advice, based on the Game Theory Optimal model.

Old vs New AI: Facing an all-in by UTG



The hand: hero called, villain had AcKc. The board ran 8Q53Q and players split the pot.

In this example scenario UTG moved all-in for 32.5 BB. Hero is in the small blind with AKo.

The former AI would fold, whilst the new AI supports the call. Let's analyze why.


Evaluation from the old AI = Fold


Evaluation from the new AI = Call


How do you analyze the correct decision in these situations?

As always in poker, our first step is to calculate the pot odds.
In this situation, the pot is $0.68 and it would cost us $0.64 to call.
Therefore, it is a good call if we are winning more than 0.64/(0.64+0.68)= 48.5% of the time.

The second step is to determine our equity. We don’t know Under the Gun's hand, but we can guess his range.

Therefore, using a poker utility, we calculate the equity of our hand against his range. For the purpose of this article, I used the free software 'Poker Stove'. There is another player to consider, the Big Blind, which will slightly alter hero’s equity. However, the risk he would wake up with a better hand is very slim. Since the Ad and the Ks are already in our hand, there is only 0.5% chance for him to have AA or KK. To simplify the math, we will ignore him or assume he will fold.

The third and last step is to make our decision by comparing the alternatives. For example, consider the first player is pushing all the combos of Ace King and the pairs of Jacks or better. Against that tight range (only 3% of the hands), Poker Stove gives 0.398 in equity to AdKs. In other words, hero wins 39.8% on average, which is below the 48.5% requirement, and he should fold.

However, we find the solution by playing around with our opponents range. If villain is pushing all the combos of Ace Queens and Ace Kings, and the pairs of 9’s or above, hero would win 48.7% of the time on average against that range (the top 5.1% of the hands) which is just slightly above what we need with the pot odds. If he is pushing a larger range, the winning chances of our hand will improve and it is an easier call. If he is pushing a tighter range we don’t have enough equity for calling with that hand.

With no information on the villain, we have to guess what range we are facing on average. Some players will push tighter and we will make a mistake and lose money on average, some will push wider and we will win more against them.

My guess is that players, on average, are pushing more than 5.1% for 32BB from UTG. Therefore I agree with the new AI and would back the hero call here. However, it is a much closer decision than one might think. Hero is risking more than 32BB and according to PokerSnowie, folding would cost him only 2.62BB in equity.

Another reason that would make the call even closer is the rake, at micro-stakes especially. Rake varies from one room to another. Under the assumption of 5 cents in rake, hero would need 0.64/(0.64+0.68-0.05)=50.39%. And villain has to push wider than 6% (88+,AJs+,KQs,AQo+), for the call to be correct. I think most of the villains would have ATs or KQo in their range and it is still a call but the decision is getting, like I said, closer.

Article written by Sharp