The hockeystick identity is given by

The Correct Answer and Explanation is:
Here is the solution to the problem.
(a) Proof using Pascal’s identity
The hockey-stick identity is
∑k=0r(n+kk)=(n+r+1r)∑k=0r(kn+k)=(rn+r+1)
. This proof proceeds by repeatedly applying Pascal’s identity, which states
(mj)+(mj−1)=(m+1j)(jm)+(j−1m)=(jm+1)
.
The proof begins with the left-hand side (LHS) of the identity. The first term in the summation,
(n0)(0n)
, is equal to 1. Since
nn
is a positive integer,
(n+10)(0n+1)
is also equal to 1. Thus,
(n0)(0n)
can be replaced by
(n+10)(0n+1)
without changing the value of the sum.
The LHS becomes:
(n+10)+(n+11)+(n+22)+⋯+(n+rr)(0n+1)+(1n+1)+(2n+2)+⋯+(rn+r)
Applying Pascal’s identity to the first two terms gives
(n+10)+(n+11)=(n+21)(0n+1)+(1n+1)=(1n+2)
. The sum is now:
(n+21)+(n+22)+(n+33)+⋯+(n+rr)(1n+2)+(2n+2)+(3n+3)+⋯+(rn+r)
Again, applying Pascal’s identity to the new first two terms gives
(n+21)+(n+22)=(n+32)(1n+2)+(2n+2)=(2n+3)
. This process, known as a telescoping sum, continues. Each step combines the first two terms into a single term. After
r−1r−1
steps, the sum is reduced to two terms:
(n+rr−1)+(n+rr)(r−1n+r)+(rn+r)
A final application of Pascal’s identity yields
(n+r+1r)(rn+r+1)
, which is the right-hand side of the identity.
(b) Proof by combinatorial argument
The identity is first rewritten using the symmetry property
(NK)=(NN−K)(KN)=(N−KN)
. The original identity
∑k=0r(n+kk)=(n+r+1r)∑k=0r(kn+k)=(rn+r+1)
becomes:
∑k=0r(n+kn)=(n+r+1n+1)k=0∑r(nn+k)=(n+1n+r+1)
This form can be proven by a combinatorial argument. Consider the problem of counting the number of ways to choose a subset of
n+1n+1
objects from a set of
n+r+1n+r+1
distinct objects arranged in a line.
The right-hand side (RHS),
(n+r+1n+1)(n+1n+r+1)
, counts this number directly, by the definition of a combination.
The left-hand side (LHS) counts the same quantity by partitioning the selections into
r+1r+1
disjoint cases. The partition is based on the position of the leftmost object chosen from the line. Let the position of this object be
pp
. The remaining
nn
objects must then be selected from the
(n+r+1)−p(n+r+1)−p
objects to its right.
For a selection to be possible, there must be at least
nn
objects to the right of position
pp
. This implies
(n+r+1)−p≥n(n+r+1)−p≥n
, which simplifies to
p≤r+1p≤r+1
. The position
pp
can therefore range from
11
to
r+1r+1
.
For each case where the leftmost chosen object is at position
pp
, there are
(n+r+1−pn)(nn+r+1−p)
ways to choose the other
nn
objects. Summing over all possible values of
pp
gives the total number of ways:
∑p=1r+1(n+r+1−pn)p=1∑r+1(nn+r+1−p)
By performing a change of index with
k=r+1−pk=r+1−p
, as
pp
ranges from
11
to
r+1r+1
, the index
kk
ranges from
rr
down to
00
. The summation becomes
∑k=0r(n+kn)∑k=0r(nn+k)
, which is the LHS. Since both sides count the same set of combinations, the identity is proven.
