I am trying to solve the following problem. A person rolls 3 die, what is the probability that the sum of these die is less than or equal to 9? I have built the following inequality model
x1+x2+x3<=9 where the domain of x's are 1<=x1,x2,x3<=6.
Letting x1 = x1+1 , x2=x2+1 and x3 = x3+1 we have the new inequality 0<=x1,x2,x3<=5. Also creating a "slack variable let x4 = 9-(x1+x2+x3) and we have the new equation: x1+x2+x3+x4=6.
Now using the popular stars and bars technique we have * *** * * and have 9 positions where we can place a bar and we have 3 bars thus (9 3) = 84 possible outcomes. Now of these I would presume that the possible outcomes to subtract (as dies cannot have 0's) would be (6,0,0,0) , (0,6,0,0),(0,0,6,0) and (0,0,0,6) = a probability of 80/6^3 = 80/216. However the apparent answer in my textbook is 84 - (6,0,0,0) (0,6,0,0) and (0,0,6,0) which is 81/216. Why is it that we do not subtract the outcome of (0,0,0,6). Also is there another model I could use for this problem as this stars and bars model my teacher taught me is quite confusing.
I agree that the stars/bars approach is tricky for this problem, but it does give you a quick way to get the right answer--if your are careful!
Your mistake is that while none of the dice can be 6, the slavck variable an acceptably be 6. So you don't need to subtract off $(0,0,0,6)$.
The easiest approach is to cleverly split the three dice into a pair (because you know probabilities for a pair of dice) and one other. To roll a $9$ or less:
And the total probability is $\frac{6+12+15+16+15+12+5}{216}=\frac{81}{216}=\frac{3}{8}$.
Another easy approach (assuming you don't know about multinomial coefficients) is to enumerate how you can roll 9 or less. Out of 216 possible rolls:
BTW, the number of ways to roll 3,4, up to 18 are $$(1,3,6,10,15,21,25,27,27,25,21,15,10,6,3,1)$$
And with 4 dice, the number of ways to roll 4 through 24 are $$(1,4,10,20,35,56,80,104,125,140,146,140,125,104,80,56,35,20,10,4,1)$$