Probability of 4 of a kind in 7 cards, if first dealt 2 different cards?

1.4k Views Asked by At

Question: In Texas Hold 'Em, you choose your hand from among seven cards: two that are dealt to you, plus five more that are available to all players.

Suppose the two cards dealt to you are of different denominations. What is the probability that you'll be able to make four of a kind out of all seven cards available to you?


I'm trying to use Bayes conditional to solve this:

Prob (4 of a kind in 7 cards after first being dealt 2 different denominations ) = ( Prob(Being dealt 2 different denominations first (in 2 cards), given 4 of a kind in 7 cards) * Prob(4 of a kind in 7 cards) ) / Prob (Being dealt 2 different denominations first (in 2 cards))

I'm looking for clarity on how to think about this problem (and, I believe, how I should avoid overthinking it as I've spent too much time pondering "what ifs").

I can't solve "Prob (Being dealt 2 different denominations first (in 2 cards), given 4 of a kind in 7 cards)", so I'm not sure if I'm stating the possibilities properly. What if I'm dealt Queen, King, and 3 more Kings in the 5 cards. Or what if there are 4 Jacks in the 5 cards and the 4-of-a-kind is made completely in the 5? The location of the 4 of a kind (if linked with one of the initial 2 cards or not), along with how to separate the probabilities of the initial 2 cards from the 5 is confusing me.

4

There are 4 best solutions below

0
On BEST ANSWER

I would recommend doing this directly rather than using $Pr(A\mid B)=\frac{Pr(A\cap B)}{Pr(B)}$

You are dealt two cards of different ranks. Without loss of generality, suppose that the first two cards dealt to you are the Ace of spades and the King of spades.

Now... in order to have a four of a kind, among the next five cards one of the following three situations must occur:

  • Among the next five cards, the remaining three aces appear
  • Among the next five cards, the remaining three kings appear
  • Among the next five cards, four of the cards are all of the same rank (not an ace and not a king)

Note that these cases do not in fact overlap.

There are $\binom{50}{5}$ possible ways to deal the next five cards (where order doesn't matter). $\binom{47}{2}$ of those ways have all three remaining aces. Similarly $\binom{47}{2}$ have all three remaining kings.

As for counting the final case, first pick which rank it is that appears four times. There are $11$ options. Then, pick which the final remaining card is. There are $46$ options here. This gives $11\cdot 46$ such choices.

All in all, this gives the probability as:

$$\dfrac{\binom{47}{2}+\binom{47}{2}+11\cdot 46}{\binom{50}{5}}$$


If you insist on using $Pr(A\mid B)=\frac{Pr(A\cap B)}{Pr(B)}$ where $A$ represents getting a four of a kind and $B$ represents the first two cards are of different ranks, I recommend using the sample space whose elements are of the form $(\text{first card},\text{second card},\text{set of remaining cards})$, which we recognize being equiprobable so we can use counting techniques.

You will get $Pr(B)=\dfrac{52\cdot 48\cdot \binom{50}{5}}{52\cdot 51\cdot \binom{50}{5}}=\dfrac{48}{51}$

Then, in calculating $Pr(A\cap B)$, we do so similarly to the first method, except having to deal with the extra step of deciding the specific cards used in the first two slots:

$Pr(A\cap B)=\dfrac{52\cdot 48\cdot \binom{47}{2}+52\cdot 48\cdot \binom{47}{2}+52\cdot 48\cdot 11\cdot 46}{52\cdot 51\cdot \binom{50}{5}}$

Completing the arithmetic arrives at the same answer as before.

0
On

You have two ways to make quads.

3 that match a card in you hand.
There are only 2 of those.
The 2 other 47 cards.

4 that match on the board
There are 11 of those.
1 of the other 46 cards.

Possible boards 5 of 50.

$$\dfrac{2\cdot\binom{3}{3}\cdot\binom{47}{2}+11\cdot\binom{4}{4} \cdot \binom{46}{1}}{\binom{50}{5}} = 0.0012592271$$

Same as JMoravitz - give him the check. I just wanted to work through it on my own.

Note in poker one card in your hand could come into play with quads on the board.

If you make quads with 1 in your hand you are guaranteed to win. A board of 5 cannot have two trips.

2
On

Let's say the first two cards are "A" and "B". To get "four of a kind", you the next three cards must be of type "A" or of type "B". There are 50 cards left to draw from, 3 of type "A" and 3 of type "B". The probability the third card matches either "A" or "B" is 6/50= 3/25. After that we must match the third card drawn. On the fourth draw there are 49 cards left, 2 of the correct type. The probability the fourth card matches the third card is 2/49. There are then 48 cards left, just 1 of the correct type. The probability the fifth card matches the third and fourth cards is 1/48.

The probability of four of a kind given that the first two cards are different is (3/25)(2/49)(1/48).

2
On

The idea that the probability cant be counted is a bit incorrect, tho not as clean.

C1 = $\frac{52}{52}$ or $P(1)$ out of 52 possible choices

C2 = $\frac{48}{48}$ or $P(1)$ out of 52 - 4 (so that the C1 and C2 don't match)

C3 = $\frac{6}{50}$ or $P(\frac{3}{25})$ since there are 3 cards to match c1 and 3 cards to match c2

C4 = $\frac{2}{49}$ or $P(\frac{2}{49})$ since there are only 2 cards in the deck to match c3

C5 = $\frac{1}{48}$ or $P(\frac{1}{48})$ since there are only 1 card in the deck to match c4

C6 = $\frac{47}{47}$ or $P(1)$ This card don't matter and since it can be anything, a probability of 1 makes things easier

C7 = $\frac{46}{46}$ or $P(1)$ This card don't matter and since it can be anything, a probability of 1 makes things easier

Thus: $\frac{52}{52}*\frac{48}{48}*\frac{3}{25}*\frac{2}{49}*\frac{1}{48}*\frac{47}{47}*\frac{46}{46}$ $=$ $1*1*\frac{3}{25}*\frac{2}{49}*\frac{1}{48}*1*1$ $=$ $\frac{3}{25}*\frac{2}{49}*\frac{1}{48}$ = $\frac {1}{9800}$ = .000102

HOWEVER

This outcome can happen with $\binom {5}{2}$ or 10 different card locations/positions (combinations)!

So: $\frac {1}{9800}*\frac {10}{1} = \frac {1}{980}$ = .0010204...

This obviously under-counts something (by -.000239) but this is a good thing as the probability of getting 4 of a kind with out any help from the first 2 pocket cards equals????? ~.000239.

C1 = $\frac{52}{52}$ or $P(1)$ out of 52 possible choices (this card can not match the river cards)

C2 = $\frac{51}{51}$ or $P(1)$ out of 51 cards left (this card can not match the river card but could match c1)

C3 = $\frac{4}{50}$ or $P(\frac{4}{50})$ since there are 4 cards available to be used for 4 of a kind

C4 = $\frac{3}{49}$ or $P(\frac{3}{49})$ since there are only 3 cards in the deck to match c3

C5 = $\frac{2}{48}$ or $P(\frac{2}{48})$ since there are only 2 card in the deck to match c3, c4

C6 = $\frac{1}{47}$ or $P(\frac{1}{47})$ since there is only 1 card in the deck to match c3, c4, c5

C7 = $\frac{46}{46}$ or $P(1)$ This card don't matter and since it can be anything, a probability of 1 makes things easier

Thus: $\frac{52}{52}*\frac{51}{51}*\frac{4}{50}*\frac{3}{49}*\frac{2}{48}*\frac{1}{47}*\frac{46}{46}$ $=$ $1*1*\frac{4}{50}*\frac{3}{49}*\frac{2}{48}*\frac{1}{47}*1$ $=$ $\frac{4}{50}*\frac{3}{49}*\frac{2}{48}*\frac{1}{47}$ = $\frac {1}{230300}$ = .000004

Now take .000004 and multiply it by $\binom {5}{4}$ or 5 and multiply by the renaming possible denominations for the 5th card in the river.

$\frac {1}{230300}$ * 5 * 11 = $\frac {55}{230300}$ = $\frac {11}{46060}$ = .000239

Next add P(of 3 cards in the river that match one card in the pocket) + P(4 of a kind in the river) = .0010204 + .000239 = .001259.

BOOM! If it's not fun, it's not math :D

NN + APL