12

the sample space of one toss of three dice is: $\Omega = \left \{ (1,1,1), ..., (6,6,6) \right \}$
so there are $6^3 = 216$ possible outcomes.

What is the probability to obtain an outcome where the sum of its three components is equal to 11?

I've considered the possible value can assume dice without a particular position and then I have considered the permutations to include every position:
$(6,4,1), 3! = 6 \\ (6,3,2), 3! = 6 \\ (5,5,1), \frac{3!}{2!} = 3 \\ (5,4,2), 3! = 6 \\ (5,3,3), \frac{3!}{2!} = 3 \\ (4,4,3), \frac{3!}{2!} = 3$

so I've summed up obtaining $27$ and the probability would be $\frac{27}{216}$

Now, consider if I have to do this same passages for sums from 3 to 18, it is very exhausting.
So, my question is: Is there any "faster" way to do that?

Yly
  • 15,292
JB-Franco
  • 852

6 Answers6

12

The generating function approach is to ask for the coefficients of $x^{n}$ in the expansion of:

$$(x+x^2+x^3+x^4+x^5+x^6)^{3}=x^3\left(\frac{1-x^6}{1-x}\right)^3$$

Now, $$\frac{1}{(1-x)^3}=\sum_{k=0}^{\infty}\binom{k+2}{2}x^k$$

And $$(1-x^6)^3=1-3x^6+3x^{12}-x^{18}$$

So the product of these and $x^3$ has, for the coefficient $x^n$:

$$\binom{n-1}{2}-3\binom{n-7}{2}+3\binom{n-13}{2}-\binom{n-19}{2}$$

Trick is to treat $\binom{j}{2}=0$ when $j<2.$

So, for example, when $n=11$, you get $\binom{10}{2}-3\binom{4}{2}=45-18=27$.


This also gives you a hint of the value when dealing with $k$ dice. Then:

$$c_n = \sum_{i=0}^{k}(-1)^i\binom{k}{i}\binom{n-(1+6i)}{k-1}$$ where in this formula, $\binom{a}{k-1}$ is zero when $a<0.$

If you don't like generating functions, this can be proven via inclusion-exclusion.


There's another approach that is a little faster for computing all values.

If $(1+x+x^2+\cdots+x^5)^3=a_0+a_1x+\cdots+a_nx^n+\cdots$ then you get that:

$$(a_0+a_1x+\cdots+a_nx^n+\cdots)(1-3x+3x^2-x^3)=(1-x^6)^3=1-3x^6+3x^{12}-x^{18}$$

What this means is that (setting $a_n=0$ when $n<0$: $$a_{n}-3a_{n-1}+3a_{n-2}-a_{n-3}=\begin{cases} 1&n=0\\ -3&n=6\\ 3&n=12\\ -1&n=18\\ 0&\text{otherwise} \end{cases}$$

or:

$$a_{n}=3\left(a_{n-1}-a_{n-2}\right)+a_{n-3}+\begin{cases} (-1)^k\binom{3}{k}&n=6k\\ 0&\text{otherwise} \end{cases}$$

Then the final coefficient of $x^n$, after we multiply by $x^3$ again, is $a_{n-3}$.

So you get:

$$\begin{align}a_0&=0+(-1)^{0}\binom{3}{0}=1\\ a_1&=3\left( a_0-a_{-1}\right)=3\\ a_2&=3\left( a_1 - a_0\right)=6\\ a_3&=3\left(a_2 - a_1\right)+a_0=10\\ a_4&=3\left(a_3-a_2\right)+a_1=15\\ a_5&=3\left(a_4-a_3\right)+a_2=21\\ a_6&=3\left(a_5-a_4\right)+a_3+(-1)^{1}\binom{3}{1}=25\\ &\cdots \end{align}$$

There's a special trick that can be applied here: $a_{n}=3a_{n-1}-3a_{n-2}+a_{n-3}$ is known to be a quadratic polynomial in $n$. So, when $n$ is not a multiple of $6$, you get that:

$$a_{n-2}-a_{n-3},a_{n-1}-a_{n-2},a_{n}-a_{n-1}$$ must be an arithmetic progression for any $n$ not a multiple of $6$.

So we have that $a_5-a_4 = 6, a_6-a_5=4$ and thus $a_7-a_6=2$, or $a_7=a_6+2=27.$ $a_8=27+0=27, a_9=27-2=25,a_{10}=25-4=21,a_{11}=21-6=15,a_{12}=15-8+(-1)^2\binom{3}{2}=10$.

Thomas Andrews
  • 177,126
  • Good answer ... shows the principle of how it can be calculated for larger parameters ... typo in the second equation ? – Donald Splutterwit Aug 16 '17 at 19:40
  • 1
    @DonaldSplutterwit Yes, typo. Fixed. – Thomas Andrews Aug 16 '17 at 19:41
  • @ThomasAndrews There's a minor error in your general formula - you only need to sum up to i=k-1 (inclusive) as given n<=6k (you can't roll a total of more than all 6s), otherwise n−(1+6i) (the top half of the binomial function) ends up negative for all possible values of n at i=k. So the case of i=k always ends up zero.

    The limit is actually more complex as once k>6 you don't need the k-1 term either, as for k=7 and n=6k (i.e. rolling all 6s) \binom{n-(1+6i)}{k-1} ends up as \binom{6k-(1+6k-6)}{k-1}=\binom{5}{6} which is zero, and that holds for larger k. But k-1 is simpler

    – Thief Aug 19 '22 at 15:47
  • 1
    Well, my formula gives $0$ when $n>6k$ too. It is not an "error," but it might be unnecessary in most usual cases. @Thief And binomials $\binom ab$ are assumed to be zero in the above formula if $a<0$ or $b<0$ or $b>a.$ These are "combinatorial binomials" - counting rather than coefficients. – Thomas Andrews Aug 19 '22 at 15:50
  • Going further, you actually need significantly less terms if you take advantage of the probability reflecting around the average value. In my calculations, you can accurately calculate up to k=7 with only summing i=0...2 for values of n <= (k*3.5) (the average of k dice). This is useful if you want to actually calculate the thing – Thief Aug 19 '22 at 16:00
  • 1
    Sure, I wasn't trying to figure out the ideal computation formula, just that a closed formula exists. @Thief – Thomas Andrews Aug 19 '22 at 16:04
  • Sure. And that fact blew my mind, let alone that it's so surprisingly simple. I was just showing it can be "optimised" to sum over a smaller range in the common case. – Thief Aug 19 '22 at 16:04
7

For the favorable cases: Since the dice can assume integer values between $1$ and $6$, we wish to find the number of solutions in the positive integers of the equation $$x_1 + x_2 + x_3 = 11 \tag{1}$$ subject to the restrictions that $x_k \leq 6$ for $1 \leq k \le 3$.

A particular solution of equation 1 corresponds to the placement of two addition signs in the ten spaces between successive ones in a row of eleven ones. For instance, $$1 1 1 1 1 + 1 1 + 1 1 1 1$$ corresponds to the solution $x_1 = 5$, $x_2 = 2$, and $x_3 = 4$. Hence, the number of solutions of equation 1 in the positive integers is the number of ways we can select two of the ten spaces between successive ones in a row of eleven ones in which to place addition signs, which is $$\binom{10}{2}$$

More generally, the equation $$x_1 + x_2 + \cdots + x_k = n$$ has $$\binom{n - 1}{k - 1}$$ solutions in the positive integers since we must choose which $k - 1$ of the $n - 1$ spaces between successive ones in a row of $n$ ones will be filled with addition signs.

However, we have counted solutions in which one of the variables exceeds $6$. Notice that it is not possible for two of the variables to exceed $6$ simultaneously since $2 \cdot 7 = 14 > 11$.

Suppose $x_1 > 6$. Let $x_1' = x_1 - 6$. Then $x_1'$ is a positive integer. Substituting $x_1' + 6$ for $x_1$ in equation 1 yields \begin{align*} x_1' + 6 + x_2 + x_3 & = 11\\ x_1' + x_2 + x_3 & = 5 \tag{2} \end{align*} Equation 2 is an equation in the positive integers with $$\binom{5 - 1}{3 - 1} = \binom{4}{2}$$ solutions. By symmetry, there are an equal number of solutions in which $x_2 > 6$ or $x_3 > 6$. Hence, the number of solutions we must exclude is $$\binom{3}{1}\binom{4}{2}$$

Hence, the number of permissible solutions is $$\binom{10}{2} - \binom{3}{1}\binom{4}{2}$$

Addendum: You asked about the sums from $3$ to $18$. If $3 \leq n \leq 8$, then the number of solutions of the equation
$$x_1 + x_2 + x_3 = n \tag{3}$$ in the positive integers with $x_k \leq 6$ for $1 \leq k \leq 3$ is $$\binom{n - 1}{3 - 1} = \binom{n - 1}{2}$$ since it is not possible for one of the variables to exceed $6$.

Notice also that by symmetry, the number of solutions with sum $3$ (all ones) is equal to the number of solutions with sum $18$ (all sixes), the number of solutions with sum $4$ (two ones and a two) is equal to the number of solutions with sum $17$ (two fives and a six), and so forth. Hence, knowing the number of solutions for $3 \leq n \leq 8$ also tells us the number of solutions for $13 \leq n \leq 18$.

The following argument shows that the number of solutions with sum $n$ is equal to the number of solutions with sum $21 - n$. If $y_k = 7 - x_k$, then $1 \leq x_k \leq 6 \implies 1 \leq y_k \leq 6$. Moreover, substituting $7 - y_k$ for $x_k$ for $1 \leq k \leq 3$ in equation 3 yields \begin{align*} 7 - y_1 + 7 - y_2 + 7 - y_3 & = n\\ 21 - y_1 - y_2 - y_3 & = n\\ -21 + y_1 + y_2 + y_3 & = -n\\ y_1 + y_2 + y_3 & = 21 - n \end{align*}

By an argument similar to that given above for $n = 11$, the number of solutions of the equation $$x_1 + x_2 + x_3 = 9$$ in positive integers not exceeding six is $$\binom{8}{2} - \binom{3}{1}\binom{2}{2} = \binom{8}{2} - \binom{3}{1}$$ By symmetry, this is also the number of solutions for $n = 12$.

Also, by symmetry, the number of solutions for $n = 10$ is equal to the number of solutions for $n = 11$.

N. F. Taussig
  • 76,571
2

It is the coefficient of $x^{11}$ in $(x(1+x+x^2+x^3+x^4+x^5))^3$ which is indeed $\color{red}{27}$ but any which a ways you solve it will be a bit of a slog.

enter image description here

Donald Splutterwit
  • 36,613
  • 2
  • 26
  • 73
2

If you want a relatively efficient pencil-and-paper solution, I recommend Thomas Andrew's answer.

If you just want to avoid being exhausted by the calculations, and you are willing to use some software to help you, you can easily get the probabilities with the help of a spreadsheet. Rather than replicate answers to other questions, let me just refer you to this answer, in which I gave a spreadsheet for twenty-sided dice, and this answer, in which I adapted the spreadsheet to work for six-sided dice.

Of course if you really just want a quick answer you can ask Wolfram Alpha. The spreadsheet, however, has the slight advantage (in my opinion) that you can examine it to see how the probabilities are derived.

David K
  • 98,388
2

Noodling out.

If the first die is $a $ then we need that the second die and the third die add to $11-a$ which means the second die is between $\max (1,11-a-6)=\max (1,5-a) $ and $\min (6,11-a-1)=\min (6,10-a) $.

So the probability is $\sum\limits_{a=1}^6\sum\limits _{b=\max (1,5-a)}^{\min (6,10-a)}\frac 16*\frac 16*\frac 16$

$=\frac 1 {216}\sum_{a=1}^6 \min (6,10-a)-\max (1,5-a)+1=\sum \min (6,10-a)-\max (0,4-a)=$

$=\frac 1 {216}(3+4+5+6+5+4)=\frac {27}{216} $.

Not sure that is easier or quicker. The max and min values make it hard to generalize.

fleablood
  • 124,253
0

Hint Show that the number of combinations with sum $n$ is just the coefficient of $x^n$ in $$(x+x^2+x^2+x^4+x^5+x^6)^3$$

N. S.
  • 132,525