I've been trying to understand this but my Maths isn't strong enough. Someone else asked a similar question regarding six-sided dice. I understand this is a binomial distribution and can fill in the formula but can't complete it.

I am trying to figure out the probabilities of success for a roleplaying game, namely the new Vampire the Masquerade 5th edition. To succeed you roll ten-sided dice and need a number of successes (six or more on a die). To attack someone you roll your dice and try to get a number of successes equal or greater than half of theirs.

So I have $10$ dice as do they, we are equal. I need to roll $6$ or more on five of my ten dice.

Thank you so much for any help.

  • $\begingroup$ One way to answer the question is to add the probabilities that exactly six, exactly seven, exactly eight, exactly nine, or exactly ten of the dice display a number that is six or greater. $\endgroup$ – N. F. Taussig Jun 11 '18 at 20:20
  • $\begingroup$ is it exactly five or five at least? $\endgroup$ – narek Bojikian Jun 11 '18 at 20:24
  • $\begingroup$ When you say "get a number of successes equal or greater than half of theirs", do you mean they also roll 10D10 and count their successes? $\endgroup$ – awkward Jun 11 '18 at 20:31
  • $\begingroup$ Sorry if I wasn't that clear. To get a you need a six or more on a ten sided dice, higher than five or 6+. $\endgroup$ – James Firminger Jun 11 '18 at 20:50
  • $\begingroup$ The rules say you can half their total dice pool and just consider that the number of successes you need to roll. So to beat a person with 10 dice I would need to to roll 6 or more on five dice. $\endgroup$ – James Firminger Jun 11 '18 at 20:51

Since you didn't mention if it's exactly or at least 5 times, I will assume in my answer that it's exactly 5 dice getting a number higher than 5 and at the end I will discuss how would the "at least" change the answer:

So reformulating the question, we have ten dice, we want for five of them to get a number between 6 and ten (which are half of all possibilities for each) and for the other five to get a number between 1 and 5 (which is also half of the possibilities).

So we only have to choose the five elements which will have our values (regardless of the order) which means we will choose 5 elements out of 10 and for each we have the probability 1/2 to get the right value, and for the other five each will have the also 1/2 probability to get the it's right value.

Hence the answer is $${10\choose5} \cdot (1/2)^5 \cdot (1/2)^5 = \frac{10!}{5!\cdot5!}\cdot \left(\frac{1}{2}\right)^{10} = \frac{252}{1024} = \frac{63}{256}$$

So back to the case of "at least", all we need to do is to sum over all the ways of getting 5, 6, 7, 8, 9 and 10 dice with a number greater than 5. It'd make it easier, if we noticed that getting 6, 7, 8, 9 and 10 dice with a number greater than 5 is a exactly the half of all the possibilities which mean it's probability is $\frac{1}{2}$, and hence the total answer would be $$\frac{1}{2} + \frac{63}{256} = \frac{63 + 128}{256} = \frac{191}{256}$$

  • 1
    $\begingroup$ Thank you so much. Yes it was at least five, I should have specified. I'm trying to create a chart of the probabilities of succeeding with different numbers of dice and differemt target numbers. Eg needing to get three (6+)s with five dice or two (6+) with four. I don't understand the section with the exclamation marks, could you perhaps explain to me what you are doing there. When I tried to work this through before I posted here, I got to 1/1024. $\endgroup$ – James Firminger Jun 11 '18 at 21:11
  • $\begingroup$ Okay sorry I though you knew what that is, n! is called n faculty, it's the product of numbers from 1 to n, and it is the number permutations of the numbers from 1 to n, ie. the number of ways to represent the numbers between 1 and n in different orders. for example: 3! = 1*2*3 = 6 and we have 6 permutation of 6 numbers which are 1,2,3 1,3,2 2,1,3 2,3,1 3,1,2 3,2,1 $\endgroup$ – narek Bojikian Jun 11 '18 at 21:32
  • $\begingroup$ however, choosing k elements out of n and the order matters would be n*(n-1).. *(n-k+1) = n!/(n-k)! since we have n methods to choose the first number, n-1 methods to choose the second, .. n-k+1 to choose the kth $\endgroup$ – narek Bojikian Jun 11 '18 at 21:34
  • $\begingroup$ the last part would be the ${n}\choose{k}$ which is the number of ways to choose k elements out of n and the order doesn't matter the value of which is n/(k! * (n-k)!) you can see it in the way that, we are choosing k elements out of n which is n!/(n-k)! but in this case we would habe considered different orderings of the same set as different choices and we have k! of each (permuting over k elements) that's why we divide over k! in addition. $\endgroup$ – narek Bojikian Jun 11 '18 at 21:36
  • $\begingroup$ In your place however, I would consider reading a basics book about combinatorics before starting doing problems (a high-school book should do in this case you only need the definitions and the intuitions behind them at first). $\endgroup$ – narek Bojikian Jun 11 '18 at 21:37

Since the actual value of the roll doesn't matter as long as we know whether it's value is greater than 5, we can simplify this problem by denoting E as the event that the roll is greater than 5. This has probability 1/2 assuming a fair dice is being used.

Now, letting z be the number of times E is true, we see that the pgf for one roll of the die is F(z) = ${1\over 2} + {z\over 2}$. Therefore, the pgf after 10 rolls of the die is ${F(z)}^{10}$ = $({1\over 2} + {z\over 2})^{10}$ = $\sum_{k=0}^{10}{10 \choose k}{1\over 2}^{10}z^k$.

In your comment, you mentioned that to win, you need to roll a 5 or more on 5 or dice, therefore, this probability is simply the tail of F(z), $\sum_{k=5}^{10}{10 \choose k}{1\over 2}^{10}z^k$ = $\frac{2^{10}+{10\choose 5}}{2^{11}}$ $\approx{.62}$


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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