All in Poker Strategy - Rules to Live By

Its a common question, but one that often requires a complex answer. Should I go all-in? Let's start by establishing some basic ground rules, then look a little deeper into the maths behind commitment decisions.

For those of us who would like a crash course in understanding when we should stack preflop, it doesn't need to be that difficult.

Assuming a six-max game with 100bb stacks, we should feel pretty comfortable stacking the following ranges by default.

Naturally we can go wider than these ranges in various situations, but by default they should be a pretty good guide. The chance of making any big mistakes with these specific ranges is very low.

This rule has been around for very many years, and helps us to understand at which point we have invested so many chips that it would be incorrect to fold.

In other words, once our opponent puts us in a situation where we need to invest more than 1/3rd of our effective stack, we should automatically be deciding whether we want to commit our entire stack. If we find the answer to this question is “no” then we should typically be folding to that initial bet.

It's nearly always better that we are the player putting our opponent to a commitment decision rather than the opposite scenario. This is because we have fold-equity. A situation where we have no fold equity is never going to be as profitable.

Imagine we are playing a 6max game with 50bb stacks and the BTN decides to shove all-in. We have no history and hold AKo in the small-blind. It's possibly OK to call in this situation, but it's really not going to be overly profitable for us, despite the fact that we have part of our standard stacking range in this situation.In short, we don't know if calling will be +EV, it really depends on the player. We know for certain however that the situation will be high variance and cause our results graph to become swingy. Perhaps we have seen that our opponent actually has a VPIP of 0 over 50 hands, in which case it would be completely fine to fold in this scenario.

Now let's imagine that BTN opens for a regular sizing, we 3bet, he 4bets, and we shove over the top. This is nearly always going to profitable. Even if our opponent ends up having AA, it's important to factor in the idea that we had fold-equity with both our 3bet and our 5bet against his overall range. It's typically this fold-equity that makes AKo an extremely profitable stack-off and we should always be more cautious in situations where we are calling a jam with no fold-equity.

We can apply this logic to draws to help us understand when shoving is correct vs playing our draws slowly.

UTG (50bb) 

MP (50bb)

CO (50bb) 

SB (50bb)

BB (50bb)

Flop (7.5bb)

So first note that if we call here, we would have invested 18bb out of our 50bb stack. Why is this relevant? We'd have invested over 1/3rd of our stack and crossed the commitment threshold. According to the commitment threshold concept, we need to make a decision regarding whether it's correct to go all-in here of whether we'll do better by just calling. Assuming we just called there would be a pot of 37.5bb on the turn and there would be 32bb remaining in the effective stacks. If our opponent shoves the turn for 32bb, we'd need to hit on the river about 31% of the time in order for a profitable turn call. Assuming only our flush-draw outs are live we would hit on the river about 18% of the time, meaning we don't have a profitable call.

We should clearly be able to see then that in this scenario it might be correct to invest over 1/3rd of our stack and still fold on the turn when we are priced out. It's important to note therefore that while the commitment threshold of 33% of the effective stacks is a commonly cited rule, it's simply a rough guide and doesn't actually work in many commitment situations.

If we were the check-raiser here with A9 we'd have to call the remaining 35bb into a pot of about 101.5bb which means we'd need to hit around 34% of the time for a profitable call. There is a chance in this situation we'd be committed, since we still have two cards to come and will connect around 36% of the time by the river. So we can see the commitment threshold concept works in some situations but not in others.

Going back to the situation at hand, since we would not be committed against a turn shove with our flush-draw, does this mean it might be better for us to directly jam over our opponents flop check/raise? Potentially, but not necessarily. So what does it depend on? Again it all comes down to fold equity. Let's assume we have roughly 36% equity here in the pot, but that our opponent is never folding. What would our expected value look like?

P (winning * amount-won) – P (losing * amount lost)   (If this makes no sense, check out the articles on Ev calculations here)

Let's imagine the same scenario now, but our opponent will fold 50% of the time to our jam. We already know the expectation of getting called (-4.38bb) and can use this to help us in the following calculation