Browse · MathNet
PrintBxMO Team Selection Test
Netherlands counting and probability
Problem
Let be an integer. Ruben takes a test with questions. Each question on this test is worth a different number of points. The first question is worth point, the second question , the third and so on until the last question which is worth points. Each question can be answered either correctly or incorrectly. So an answer for a question can either be awarded all, or none of the points the question is worth. Let be the number of ways he can take the test so that the number of points awarded equals the number of questions he answered incorrectly. Do there exist infinitely many pairs with and ?
Solution
For the first few values of , note that , , , , . We claim that for , is strictly increasing as a function of . Therefore there is only a finite number of pairs as in the problem.
We view a way of taking a test as a subset of by taking the set of numbers of questions that are answered correctly. We say that a subset of is an -equally correct set if the sum of all elements of is equal to the number of elements in the complement of in . So by definition, is the number of -equally correct sets. Note that a subset of is -equally correct if and only if the sum of all elements of plus the number of elements of equals .
We first claim that is non-decreasing for . Let be an -equally correct set. Then adding to the largest element of gets us a subset of . This subset has as many elements as and has sum higher than the sum of , which is therefore an -equally correct set. This procedure defines an injective map from the set of -equally correct sets to the set of -equally correct sets for all . Therefore for all .
We now show that, for , there exists an -equally correct set that is not in the image of . Note that an -equally correct set of which the two largest elements differ by exactly cannot be in the image of ; the reason is that a set in this image is one that is obtained by adding to the largest element of a subset of . If with , then is -equally correct, as . Similarly, if with , then is -equally correct, because . We conclude that for all .
We view a way of taking a test as a subset of by taking the set of numbers of questions that are answered correctly. We say that a subset of is an -equally correct set if the sum of all elements of is equal to the number of elements in the complement of in . So by definition, is the number of -equally correct sets. Note that a subset of is -equally correct if and only if the sum of all elements of plus the number of elements of equals .
We first claim that is non-decreasing for . Let be an -equally correct set. Then adding to the largest element of gets us a subset of . This subset has as many elements as and has sum higher than the sum of , which is therefore an -equally correct set. This procedure defines an injective map from the set of -equally correct sets to the set of -equally correct sets for all . Therefore for all .
We now show that, for , there exists an -equally correct set that is not in the image of . Note that an -equally correct set of which the two largest elements differ by exactly cannot be in the image of ; the reason is that a set in this image is one that is obtained by adding to the largest element of a subset of . If with , then is -equally correct, as . Similarly, if with , then is -equally correct, because . We conclude that for all .
Final answer
No; there are only finitely many such pairs.
Techniques
Recursion, bijection