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.

Why use small bet sizes in poker

Why are we using smaller bet sizes and when do they maximize our value?  If this is a question you have about the new PokerSnowie AI you’re probably not alone.  To answer this question let’s consider why we use different bet sizes in the first place. 

When to Bet Large

A large bet is great when our range is strong and we are going to have a lot of high equity hands that can be balanced out by low equity hands (bluffs). I would call this a 'polarized range' and the more polarized our range, the larger we should bet. I also believe that we can bet larger when our range has significantly more equity.

Here is an example:
We are in the Cutoff and open half pot; it folds to the Big Blind who calls. The flop is AdKs7s and our opponent checks.

Using PokerSnowie ranges, our preflop equities are shown below.

PokeSnowie situation

Our range is significantly ahead of our opponent's here and he has many hands which shouldn’t fold to a pot sized bet. Many of the hands that will call are still losing to our range. Think about all the one pair hands our opponent holds here that cannot fold to a flop bet.
Looking at the heat map below we can see that our large bet is putting a lot of pressure on the middle part of our opponent’s range (the yellow hands). 

So what do we do when presented with a situation like this?  We bet big! PokerSnowie thinks so too.

Basically, when we have a strong range and can get value from so many of our opponent’s medium strength hands, we want to get as much money in as possible in the pot.

When to Check

Now that we have an idea of when betting big is correct, let's look at a situation where we would usually check. What would be a high-frequency situation for checking? Based on when we bet big, I would say it is probably a situation where our overall equity is slightly lower, but more importantly, our equity distribution is worse… Example time!

We again open with half pot in the Cutoff and our opponent in the Big Blind calls. The flop is Ac6d5s and our opponent checks.

PokerSnowie situation

Think about our range here and how many hands we have that are worse than a top pair or a weak top pair that won't have worse hands calling. Weak aces, pocket pairs, even king high has equity here. If we make a large bet with a pair of kings we can expect to lose most of the time when called, even though KK has almost 70% equity it is a clear check in this spot.

A bet in this spot, especially a large one, will force many of our opponent’s hands to fold, which is easy for him since he has so many low equity hands.  Here is what PokerSnowie thinks:

PokerSnowie advice

Notice that PokerSnowie is checking here about 60% of the time. This is a high frequency in this position. One thing we might note is how many medium and weak equity hands our opponent has.  If we bet large here it's reasonable to assume our opponent would want to get away from most of these hands.

Would it be correct to call with 53s against a pot size bet?  Probably not, so how can we put pressure on our opponent without letting him easily play perfectly against us?

By betting small!

Something to note is that the checking frequency would have been even larger using the old AI.  The smaller bet size will allow us to bet more frequently and therefore extract more value from the middle of our opponent’s range. Notice that even though our opponent has a lot of medium equity hands, they are so weak that betting small or checking works best.

When to Bet Small

In the last example, we look at a clear spot where a small bet is superior.

We again open to half pot in the Cutoff and our opponent in the Big Blind calls.  The flop is QsQh7s and our opponent checks.

PokerSnowie situation

Just like the example where we check frequently, the more marginal our equity distribution, the more we should think about betting small.  On this board we won’t have a strong hand very often, and when we do it means our opponent almost certainly doesn’t.  So betting big will only get called by very strong hands.  We can manipulate our opponent’s calling range by changing our bet size.

In a $100 pot, when we bet $100, our opponent must call 50% of the time to keep us from making an instant profit1.

1- [100 ÷ (100 + 100)] = .50

If we bet $25 into that $100 pot, however, he must defend 80% of the time to keep us from making an instant profit.

1- [25 ÷ (25 + 100)] = .80

This means our opponent should call with 80% of his range. All of a sudden we can get a lot more value from his range with both our strong hands and our medium strength hands.  It maximizes our value against those weak hands that a larger bet would allow him to fold.

Let’s see what PokerSnowie thinks!

PokerSnowie advice

Lastly, we will look at the equities of the ranges in heat map form and see how betting small manipulates our opponent into calling more.

When we can bet a balanced range, we make money, so being able to bet more frequently and force our opponents to defend with small EV hands will make more money in the long run than having them fold big negative EV hands!

I hope you enjoyed my first article for PokerSnowie!

In my next articles, we will look at specific hands where the smaller bet size is now chosen and whether or not it makes sense based on the theory discussed in this post.

Article by Eric Graul 


This is oversimplified, it does not consider the price preflop or the fact that an opponent can raise sometimes.  Raising decreases the total required defending frequency.  This still accurately demonstrates the point I am trying to make.