Knights Around A Table

25 of King Arthur's knights are seated at the round table. Three of them are randomly chosen to be sent off to slay a troublesome dragon. Let P P be the probability that at least two of the three had been sitting next to each other. If P P can be expressed as a b \frac {a}{b} , where a a and b b are pairwise coprime integers, find a + b a+b .


The answer is 57.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

7 solutions

Michael Mendrin
May 29, 2015

Let's say the 3 knights picked in order are named A, B, C as distinct. For symmetry reasons, let A be at seat 1. Then there are (24)(23) ways to arrange B and C in the rest of the seats. Out of these, we count the distinct ways at least 2 of them are sitting next to each other. Either B or C can be seated next to A, in which case the other can be seated in 23 other places, making (4)(23) in all. If neither B or C is seated next to A, then they must sit next to each other, and they can be seated in 20 other places times 2 for different orders, making (2)(20) in all. The fraction we seek is (4)(23) + (2)(20) divided by (24)(23), which works out to 11/46, and so the answer is 57.

Nice solution. When I read "at least" I immediately decided to count the complement of none sitting together. I found it easier than trying to figure out a way to directly count ways in which at least 2 sit together.

Kunal Kantaria - 6 years ago

Could someone tell me why my reasoning is not correct?

1st person sits at the table, 24 seats left. 2nd person has 2/24 chance to be seated next to him. If he won't, then third person has 23 seats to choose from, 4 of them will be next to one of the previous 2 people. So 3rd person has 22/24 * 4/23 chance to sit next to any of previous two. So the probability should be 2/24 + 22/24 * 4/23 = 67/276 ~ 0.243 But it is 11/46 ~ 0.239

Can anyone tell why it is not working?

Bartosz Lulka - 1 year, 1 month ago

Log in to reply

The problem is the assumption "4 [seats] of them will be next to one of the previous 2 people". In most cases that is correct, but if the first two knights have exactly one seat between them, the third knight has only three seats to choose from, making the event slightly less common.

That's the reason why your probability is slightly bigger than the solution!

Carsten Meyer - 1 year, 1 month ago

Log in to reply

Thank you! That's what I was missing!

Bartosz Lulka - 1 year, 1 month ago

I have yet to see a PIE solution so here we go: The denominator is obviously 25c3. So we must find the number of ways we can choose consecutive knights. Lets label all 25 spots as the first 25 positive integers (1,2,3,...25). How many pairs of consecutive integers are there? (1,2),(2,3),...(25,1), 25 pairs. Now for the third number (since we are taking 3 people) there are 25-2 or 23 possibilities. 25*23=575. However notice we must consider the cases where ALL 3 numbers are consecutive. What you might notice is that we have in fact already counted that, but twice as much as we needed too. So we need to find the number of consecutive triplets from 1-25. (1,2,3), (2,3,4), (3,4,5),... (25,1,2) we have 25 of these. Applying PIE we take 575-25 and get 550 which is our numerator. 10c3 simplifies too 2300 and 550/2300=11/46 so a+b is 57.

Mr. Krabs - 2 months, 4 weeks ago

AB next to each other, C sits in other places (It should be 21 ) and times 2 for BA and we have here 3 cases for AB/C, AC/B and BC/A. so we get here 42+42+42 =42(3) ways for this case.

For another case is ABC all sit next to each other, that is 3! = 6 ways. Therefore, we have here 42(3) + 6 = 132 ways.

Wutthichai Misong - 5 days, 9 hours ago
Patrick Heebels
Jun 2, 2015

Great ideas of Michael Mendrin en Kunal Kantaria in my opinion. I hope my solution is viewed as an useful addendum and not as redundant.


Denote P r n = n ! ( n r ) ! P_r^n = \frac{n!}{(n-r)!} as the number of r r -permutations of n n distinct objects and Q r n = P r n r Q_r^n = \frac{P_r^n}{r} as the number of r r -circular permutations of n n distinct objects.


Special cases where r = n r = n give rise to P n n = n ! P_n^n = n! and Q n n = ( n 1 ) ! Q_n^n = (n-1)! .


The question: "In how many ways can three specific men (A, B and C) along with 22 other men be seated around a circular table in such a way that at least of two the three men (A, B, C) are seated next to one another?" is equivalent to the problem at hand.

The total number of ways to seat 25 men around the table is

Q 25 25 = ( 25 1 ) ! = 24 ! Q_{25}^{25} = (25-1)! = 24!

Solution 1: The method of Michael Mendrin

Case 1: The three men are seated next to one another in Q 23 23 3 ! = 22 ! 3 ! Q_{23}^{23} \cdot 3! = 22! \cdot 3! number of ways. Treat the three men as one entity and place the 23 'objects' around the table and consider the number of ways the three men can permute.

Case 2: Of the three men only two are seated next to one another can be done in

Q 23 23 21 2 3 = 22 ! 21 3 ! Q_{23}^{23} \cdot 21 \cdot 2 \cdot 3 = 22! \cdot 21 \cdot 3!

number of ways. Treat the two men as one entity and place the 22 + 1 'objects' around the table. The entity of two men can be seated at 21 places (not adjacent to the specific man), the two men can change seats with one another (factor 2) and there are 3 ways of choosing two men out of the three to form the entity.

Total number of seatings is 22 ! 3 ! + 22 ! 21 3 ! 22! \cdot 3! + 22! \cdot 21 \cdot 3! out of 24 ! 24! seating resulting in:

22 ! 3 ! + 22 ! 21 3 ! 24 ! = 22 ! 3 ! ( 1 + 21 ) 24 ! = 11 46 \frac{22! \cdot 3! + 22! \cdot 21 \cdot 3!}{24!} = \frac{22! \cdot 3! (1 + 21)}{24!} = \frac{11}{46}

Solution 2: By method of complementation of Kunal Kantaria

The complement of the number of desired seatings is the number of seatings in which none of the three men (A, B, C) are seated next to one another. The 22 other men along with man A can be seated in Q 23 23 = 22 ! Q_{23}^{23} = 22! number of ways. There are now 23 places where man B can be seated, but only 23 - 2 = 21 seats are allowed. Then 24 men are seated and of the 24 available seats only 24 - 4 = 20 seats are allowed for man C.

Thus the total number of desired seatings is 24 ! 22 ! 21 22 24! - 22! \cdot 21 \cdot 22 giving rise to the same requested probability:

1 22 ! 21 22 24 ! = 1 35 46 = 11 46 1 - \frac{22! \cdot 21 \cdot 22}{24!} = 1 - \frac{35}{46} = \frac{11}{46}

The answer in both solutions is 11 + 46 = 57.

I have to go back to school to learn this.

Michael Mendrin - 6 years ago

I think this is equivalent to finding the number of triangles formed by the vertices of a 10-sided polygon, such that the triangles have one or two side(s) common with the polygon.

Saurabh Chaturvedi - 5 years ago

I'm confused can you please explain how there are 23 places where man B can be seated. I understand how you got 21 but im confused as to how you reached the number 23. Thank You!

Ashish Sacheti - 4 years, 10 months ago

Log in to reply

Think of the 23 men currently around the table as a clock face: there are 23 spaces between them.

Richard Hickson - 3 years, 5 months ago
Laurent Shorts
Apr 21, 2016

There's 25 ways to pick up three people next to each other.

With 2 people next to each other, there's 25 ways to pick up the single one, and then 25 4 = 21 25-4=21 ways to pick up two people next to each other. So that is 25 21 = 525 25·21=525 ways.

Total: 550 ways. Total of ways is ( 25 3 ) = 2300 \begin{pmatrix}25\\3\end{pmatrix}=2300 .

Then P = 550 2300 = 11 46 P=\frac{550}{2300}=\frac{11}{46} .


With formula of number of m m -gons in a n n -gon without common side (so, picking m m people out of n n people without two next to each other) is n m ( n m 1 m 1 ) \frac{n}{m}\begin{pmatrix}n-m-1\\m-1\end{pmatrix} . Here, that is 25 3 ( 21 2 ) = 25 3 210 = 1750 \frac{25}{3}\begin{pmatrix}21\\2\end{pmatrix}=\frac{25}{3}·210=1750 , so 2300 = 1750 = 550 2300=1750=550 ways with at least two adjacent people.

I explained the formula in the answers of this problem .

Lovro Cupic
Jan 21, 2018

In 2 24 \frac{2}{24} cases, the 2nd knight is already next to 1st. In 2 24 \frac{2}{24} cases, the second knight is one chair apart from 1st, so there is a 3 23 \frac{3}{23} chance that the 3rd knight will be a neighbor of one of them. In remaining 20 24 \frac{20}{24} scenarios, the probability for 3rd knight to be a neighbor of one of them is 4 23 \frac{4}{23} , so the probability in question is simply 2 24 \frac{2}{24} + 2 24 \frac{2}{24} 3 23 \frac{3}{23} + 20 24 \frac{20}{24} 4 23 \frac{4}{23} = 11 46 \frac{11}{46} .

Shubham Garg
Jun 9, 2015

Let the three knights who are sent be A , B , C A,B,C .They can be seated around the table in 2 ways A B C ABC or A C B ACB clockwise.

Let us now find the probability Q Q that none of them is sitting together.

Let there be x , y , z x,y,z persons sitting between A B A B and C C respectively

where x , y , z x,y,z should be greater than 0 0 (not equal to 0 0 )

and x + y + z = 22 x+y+z=22

Using multinomial therem we have the number of solution as ( 21 2 ) \dbinom{21}{2} and the knights can be arranged in 22 ! 22! ways.

The total number of sitting arrangements are ( 25 1 ) ! = 24 ! (25-1)!=24! .

Q = 2 × ( 21 2 ) × 22 ! 24 ! = 35 46 \Rightarrow Q = \frac{2 \times \dbinom{21}{2} \times 22!}{24!}=\frac{35}{46}

Therefore the required probability P P that at least two are sitting together is P = 1 Q P = 1-Q

P = 1 35 46 \Rightarrow P=1-\frac{35}{46}

P = 11 46 \Rightarrow P=\frac{11}{46}

Therefore the answer is 57 \boxed{57}

Joshua Prettyman
Jun 2, 2015

Number of way we could have three knights all in a row = 25

Number of way we could have two knights next to each other and one sitting alone = 25 × 21 25\times 21 . There are 25 ways of having two seats in a row, 21 possibilities left for the choice of the third knight.

Number of ways we could have all three sitting singly = ( 25 × 20 × 19 + 25 × 2 × 20 ) / 3 ! = 25 × 70 (25\times 20\times 19 + 25\times 2\times 20)/3! = 25\times 70 . There are 25 ways to pick the first knight, 20 ways to pick the second so that there is a gap of at least two between them in which case there are 19 ways to pick the third knight. There are also 2 way to pick the second knight so that there is a gap of just one between them, in which case there are 20 ways to pick the third knight. Divide all this by 3! because we can reorder.

So we have 25 + 25 × 21 + 25 × 70 25 + 25\times 21 + 25\times 70 ways to pick the three knights and 25 + 25 × 21 25 + 25\times 21 of these give at least two sitting next to each other. get rid of the factor of 25 and we have (1+21)/(1+21+70)= 22/92 = 11/46.

Not the quickest way (we didn't really need to work out the number of ways we could have all three sitting singly), but it shows all the different combinations.

Let's look at a round table of 6 knights. It might help us to imagine it as a 6 sided polygon:

Now, let's select any pair of adjacent seats from the table. This is equivalent to selecting a random edge from the hexagon, giving us 6 possible choices. With N N seats, we have N N possible adjacent seat pairs to choose from. Next we choose a random point from the remaining points. For the hexagon, this would seem to give us 6 2 6-2 possible seat choices. So we would naively expect there to be 6 ( 6 2 ) 6*(6-2) ways to select the knights such that at least two of them are adjacent to each other. However, we have forgotten to notice that we're double counting.

Let's go through the seat-pairs for a round table of 6 knights where the seats are labeled 1-6 and 1 and 6 are adjacent:

12 3 4 5 6
23 1 4 5 6
34 1 2 5 6
45 1 2 3 6
56 1 2 3 4
61 2 3 4 5

I've bolded the seats that create duplicate selections. For example, the selection ( 1 , 2 ) + 3 (1, 2) + 3 is the same the selection ( 2 , 3 ) + 1 (2,3) + 1 . For each adjacent-pair there are going to be exactly 2 seats that they'll share with other adjacent-pairs no matter what N N we use since each adjacent-pair has exactly 2 points adjacent to the pair. So to fix our double counting issue we exclude the point directly counterclockwise from the adjacent-pair we chose:

12 3 4 5
23 4 5 6
34 1 5 6
45 1 2 6
56 1 2 3
61 2 3 4

We've eliminated one of the choices from our original N 2 N-2 and are left with N 3 N-3 ways to uniquely select a final seat for each adjacent-pair.

Thus there are N ( N 3 ) N*(N-3) ways to uniquely select the knights such that at least 2 of them are adjacent to each other.

Finally there are ( N 3 ) \begin{pmatrix} N \\ 3 \end{pmatrix} ways to select knights to go on the mission. Therefore the probability that at least 2 of the knights are adjacent to each other is: N ( N 3 ) ( N 3 ) \frac { N*(N-3) }{ \begin{pmatrix} N \\ 3 \end{pmatrix} }

And when you plug N = 25 N=25 in you get: 550 2300 \frac {550 }{ 2300 } which simplifies to 11 46 \frac {11}{46} . Finally giving us the answer 57 57 .

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...