Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

How can I calculate the probability of shaking 7 of a kind using 10 six sided dice?

share|improve this question
add comment

4 Answers

up vote 5 down vote accepted

If you want exactly $7$ of a kind, you use binomial distribution to first find the probability that you roll $7$ of the same "good" number in $10$ trials. The binomial distribution shows that this probability is $$ \binom{10}{7}\left(\frac{1}{6}\right)^7\left(\frac{5}{6}\right)^3, $$ since the probability of success is $1/6$.

Then there are six numbers to choose one from which could be the "good" kind of which you want $7$. Hence you multiply the above expression by $\binom{6}{1}$: $$ \binom{6}{1}\binom{10}{7}\left(\frac{1}{6}\right)^7\left(\frac{5}{6}\right)^3. $$

Note that this gives the exact probability $\frac{625}{419904}\approx\frac{1}{672}$, consistent with the approximate value given by Wolframalpha.

share|improve this answer
    
Your last sentence is not correct. Notice what happens if you try to use this to determine the probability of 2 of a kind, for instance. –  Matthew Conroy Jan 11 '11 at 20:44
    
@Matthew Conroy, thank you, I have removed it. Was the issue with the power of $3$? I was assuming that when changing the $7$, one would also change the $3$ to a new value, say $8$ in the case of $2$ of a kind. Is that the only issue, or am I missing something else? –  yunone Jan 11 '11 at 20:49
    
The trouble is that the (5/6)^(10-n) bit allows the "other" dice to be anything, which could include pairs, or triples, etc. So when you consider the 2 of a kind case, for instance, you would not be considering the fact that one can have more than one two of a kind at a time. Hence, when multiplying everything by 6, you will have over-counted the ways to have a two of a kind. –  Matthew Conroy Jan 11 '11 at 20:57
    
@Matthew, interesting, I hadn't thought about that. So for 5 of a kind of smaller, the problem becomes a little different? Thanks again for catching that error. –  yunone Jan 11 '11 at 21:02
add comment

Calculate the probability of exactly 7 sixes.
Calculate the probability of exactly 8 sixes.
Calculate the probability of exactly 9 sixes.
Calculate the probability of exactly 10 sixes.
Add them all up, and multiply by 6.

share|improve this answer
    
Thanks for your answer, that is helpful. One more question though. I can calculate the probability of exactly 10 sixes but how do I calculate 9 out of 10 being a certain number? –  Shaun Bowe Jan 11 '11 at 19:20
    
You need 9 dice to be 6 and 1 to be something else. There are 10 ways to select which one is the something else. So the probability is $\left(\frac{1}{6}\right)^9\frac{5}{6}10$ –  Ross Millikan Jan 11 '11 at 20:58
add comment

To be short: The result is $p\approx \frac{1}{672}$

You can use the second formula in the Wikipedia article and sum the probabilities for every face of 1-6: http://en.wikipedia.org/wiki/Liar%27s_dice#Basic_dice_odds

Note that Wolframalpha is also able to get this: http://www.wolframalpha.com/input/?i=10+dice (Press on more under probabilities)

share|improve this answer
1  
Though the links are helpful, this answer is not correct: 1/672 is only an approximation. See yunone's correct answer. –  Matthew Conroy Jan 11 '11 at 20:28
    
Thanks for the hint, I adjusted it and agree that yunones answer is better. Still the formula he mentions is basically the one that I linked on Wikipedia :-) Stupid me in trusting that wolframalpha calculates it accurately –  Listing Jan 11 '11 at 20:33
    
Note all the "approximately equal" signs that WolframAlpha uses on that page. For instance, 0.2718 is clearly not 1/4. –  Matthew Conroy Jan 11 '11 at 20:40
    
Yes, I thought the left is the approximation of the exact fraction on the right side not looking at the small examples. –  Listing Jan 11 '11 at 20:45
add comment

Here's another way to get the same answer (and let's be explicit that we're looking for the probability of exactly 7 of a kind, not at least 7 of a kind):

Write the result as a sequence, e.g., the ten rolls might come out $(2,5,4,6,5,4,3,1,2,2)$. Obviously there are $6^{10}$ such sequences, each equally likely.

How many sequences have 7 of a kind? There are 6 possibilities for what we have 7 of, and there are $10\choose 7$ ways to pick the 7 places that are the same. Finally, there are $5^3$ ways to fill the other three places, since there are 5 choices for each place.

So the answer is $6\cdot{10\choose 7}\cdot 5^3 / 6^{10} = 625/419904$.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.