Probabilities of repeated card draws

549 Views Asked by At

I draw from a deck of 40 cards 9 cards (edit: without replacing them). 3 of the cards in the decks are a Joker. What is the chance I drew at least 1 Joker?

My initial thought was: "Easy!, the first draw chance for a Joker is 3/40, the chance for a Joker on the second draft is 3/39, etcetera." But if I just add up all this chances, something goes wrong. None of those chances is smaller than 3/40, so drawing 14 cards would add up to more than 42/40... above 100%. Experience says, that the chance to draw 15 or more cards out of 40 without drawing one of 3 Jokers is definitely non-zero. The chance will only be zero after drawing 38 cards, as all 3 Jokers could randomly end up on the bottom of the deck.

I believe (and find it hard to exactly explain why) that I have to multiply each chance for a Joker per draft with the chance, that I did NOT draw a Joker in a previous draft, before I add the chances up. So, for 3 drafts the chance for a Joker is

  • (3/40) <-first draft
  • +(1-(3/40)) *(3/39) <-the chance first draft failed, times second draft)
  • +((1-(3/40)) +(1-(1-(3/40))*(3/39)) *(3/38) <-the combined chance the first two drafts failed, times third draft)

which makes for quite horribly long and complicated equations, if I want to calculate the chances for more drafts.

I also believe that I can simplify the calculation, by instead simply calculating the overall chance to NOT have drawn a Joker, which I then can simply deduct from 100% to get the chance of drawing a Joker.

So, if I am correct in my assumption, the chance to NOT draw a Joker at first draft is 37/40. On second draft, the chance to NOT draw a joker is 36/39, so the total chance would be (37/40) *(36/39) to NOT have drawn a joker, third chance would be (37/40) *(36/39) *(35/38) etcetera. Is that correct? Let's check with a deck size of 5, 3 Jokers and 3 drafts. (2/5) *(1/5) *(0/5)=0. The chance to not have drawn one of 3 Jokers from 5 cards after 3 drafts is 0. That seems correct.

Is there a way to generalize this equation with n as the final probability to NOT have drawn a Joker, s as the deck size, j for number of Jokers in deck and d for number of drafts? I assume it should be possible by using integral and sigma signs, but I am not very practiced at using them. Or maybe by using a minimum function?

And, I would be very thankful, if someone can point me to a free online calculator, which accepts that exact generalized equation, resp. lets me enter it with a finite number of clicks, even for a lot of drafts, so I can finally get to optimizing my deck strategies for collectible card games?

1

There are 1 best solutions below

2
On

From your attempts, the draws are without replacement.

Your formula towards the end is also correct, but the same result can be obtained in a way that you may find easier, using the formula for combinations, for which combination calculators are easily available online.

Out of $40$ cards, $3$ are jokers, $37$ non-jokers, and the formula is

$Pr = \;^{37}C_5 /^{40}C_5$

which is the same as $\frac{37}{40}\frac{36}{39}\frac{35}{38}\frac{34}{37}\frac{33}{36}$

PS The above formulas are for not drawing any joker, that for drawing at least one joker would be the complement of the above, viz using the combinations formula,

$ 1 - [^{37}C_5 /^{40}C_5]$