Now that we understand a little more about Omaha 8/b, assuming you read the last article, let’s look a little more about the math behind a split pot and the odds you need to play specific types of hands.

To start, in Hold’em, there are monster starting hands that are prohibitive favorites against any other kind of hand. AA is no worse than a 75% favorite against any other two hole cards. In Omaha 8/b, the best starting hand is AA23 double-suited, because all of your cards combine to make nut high and nut low hands. Compared to the worst of Omaha hands, uncoordinated rags like Q963 rainbow, AA23 is no better than 2-to-1 to take the high. It’s considerably better for the low….but…

Not every hand will have both a high and a low…for there to be a low, the board must have three separate cards ranked 8 or lower. First, knowing that they have to be of separate rank is important. A board that shows 2233K cannot make a low, because you have to use three of them, and splitting the pairs only gives us a K32. The odds of the board making a low from any five cards in the deck are 61.7% (the math follows at the end, for those who are curious).

So, why should you care? Well, the goal of a split-pot game like Omaha 8/b is to scoop the entire pot. As we established last time, quartering kills. If the board will play high-only less than 40% of the time, you cannot scoop enough with a high-only hand to make playing them worth your while unless you know your high possibilities will give you the nuts on a high-only board. If you’re holding K Q J 10, you need a very specific high flop that includes an A in order to make your nut straight. You are very vulnerable to a flush in any suit as well as any board that makes a full house possibility. There is also the likelihood of having to split the pot with your nut straight against any player with similar cards to yours. The bottom line is that these hands should be thrown away if you’re in a game where multiple players are seeing the majority of flops.

Since you’re playing multiple combinations of cards from your hand, the pot odds calculations you must perform are also far more complicated than those in a game like Hold’em. As an example, you’re on the button and dealt A K J 9. The flop comes 10 8 3. You have eight outs to the nut straight (a Q or 7) and seven more outs to the nut flush (any other spade), which means that normally you’d want to call any bets that come your way. However, 23 of the 45 cards left unseen will also make the possibility of a low hand…over half the deck will prevent you from scooping the pot, meaning that your implied odds on the hand are drastically different. Add into that consideration that if the board pairs, the full house should give you pause as well. What do you do when the turn is the 10? Those sorts of considerations have to be taken into the equation when you look at your turn to act when the first player bets the flop, the second calls, and the third raises…and now it’s your turn to act, and that beep on your computer tells you there’s only 10 seconds left to come up with the right decision.

I won’t try to address specific answers to any of those issues I just raised. Hold’em, especially limit hold’em, is often times a game of hard and fast answers. Omaha, especially 8/b, is often times purely a game of feel, image, intuition, and yes, luck. There’s too many combinations out there to try and determine exactly what you should do every time.

Omaha 8/b is a game of draws and possibilities. In Hold’em, the best hand on the flop is usually a considerable favorite to win the hand. In Omaha, that’s almost never the case. Learning how to play those situations are both the most challenging and most interesting part of the game, and the most important to making you a profitable player.

Appendix A: The Math of a Low Hand

In probability theory, the odds of a hand with probability P are expressed as 1/P. So, let’s calculate the P for a scenario where three cards of the five on the board are 8 or lower and none of those three are paired.

There are 32 cards in the deck that can play in a low hand (four each of A, 2, 3, 4, 5, 6, 7 and 8). The P of any one card coming at as a low card, then, are 32/52 = .615. The second card coming off cannot pair the first, and the third cannot pair either of the first two. The P of three consecutive cards making a low, then, are (32/52) * (28/51) * (24/50) = .162.

Now it gets tricky, because we’re not just looking for any three consecutive cards, we’re looking for a combination of any three cards out of the five on the board. There are 10 possible combinations of three cards in the five on the board (we could start talking about binomial coefficients here, but it’d start to get ugly…you can just count it out for yourself if you don’t believe me in this example). For any three cards out of the five making a low, P = .162 * 10 = 1.62. The odds, then, are 1/1.62, or approximately 61.7%.

If you have a hand with four low cards already, these numbers change slightly. Leaving the math as an exercise to the reader, the odds drop approximately 5%, depending on which combination of four low cards you use, whether they’re paired, etc. Of course, all of this math goes right out the window every time I’m holding A234…in that case, the odds of a high-only hand appearing aren’t approximately 43%. They’re far closer to 100%.