Simpler way to calculate probability of a certain card in a set of drawn cards

79 Views Asked by At

If you draw 9 cards from a 52-card deck, what is the probability of one Ace and one King?

I realize you can calculate this by using the hypergeometric distribution as:

$$ p(k=2, M=52, n=8, N=9) = \frac{\binom{8}{2} \binom{44}{7}}{\binom{52}{9}} = 0.2916 $$

However, my question is if there is an easier way to calculate this, that you would for example be able to approximate in your head without pen and paper?

1

There are 1 best solutions below

2
On

The only alternative method of computation for this problem that I know of is to (partially) forgo a Combinatorical approach in favor of a probability of events approach. However, I don't recommend the alternative approach, described below.

Let $E_1$ denote the event that a King is drawn on the first card.

Let $E_2$ denote the event that an Ace is drawn on the second card, given that a King was drawn on the first card.

Let $E_3$ denote the event that cards other than an Ace or King were drawn on the next $7$ cards, given that an Ace and a King were drawn on the first two cards.

Then, the probability of first drawing a King, then drawing an Ace, and then drawing $7$ other cards, besides an Ace or a King is :

$$T = p(E_1) \times p(E_2) \times p(E_3) = \frac{4}{52} \times \frac{4}{51} \times \frac{44 \times 43 \times \cdots \times 38}{50 \times 49 \times \cdots \times 44}. \tag1 $$

At this point in the analysis, I see no alternative to completing the computation other than (partially) resorting to Combinatorics.

In (1) above, $T$ specifically expressed the probability that the King was the very first card drawn, and the Ace was the second card drawn. In fact, from the problem's constraints, there are $9$ choices for which card drawn is the King, and then $8$ choices for which of the remaining cards drawn was the Ace.

So, the final computation would be

$$9 \times 8 \times T.$$

Personally, beside this approach being somewhat convoluted, I don't see this approach being easier to manually convert into the form $~\displaystyle \frac{m}{n}, ~: ~m,n \in \Bbb{Z^+},~$ than the original Combinatorics approach.

Edit
For what it's worth, I will take a crack at the mental conversion.

I would convert :

  • $\displaystyle (9 \times 8) \times \frac{4}{52} \times \frac{4}{51} \approx \frac{72}{169} \approx \frac{1}{2}.$

  • $\displaystyle \frac{44 \times \cdots \times 38}{50 \times \cdots \times 44} \approx \left[\frac{41}{47}\right]^7.$

Then, I would reason that when $a > b > 0$, that $~\displaystyle \left[1 - \frac{b}{a}\right]^2 \approx 1 - \frac{2b}{a}.$

So, I would very roughly approximate $\displaystyle ~\left[\frac{41}{47}\right]^7~$ as $\displaystyle \left[1 - \frac{5 \times 6}{47}\right] = \frac{17}{47}.$

Then, I would approximate

$$\frac{1}{2} \times \frac{17}{47} \approx \frac{8.5}{47} \approx \frac{2}{11}.$$

Frankly, I have no idea how close the computation of $~\displaystyle \frac{2}{11}~$ actually is to the right answer. $~\displaystyle \frac{2}{11}~$ would simply be my blind guess.