Kvant Math Problem 173
Let the magic sum be $M$.
Verified: no
Verdicts: SKIP + SKIP
Solve time: 7m12s
Source on kvant.digital
Problem
In the square array $4 \times 4$, the numbers 1, 2, 3, $\ldots$, 16 are arranged so that the sum of the four numbers in each row, in each column, and on each of the two diagonals is the same, with 1 and 16 placed in opposite corners of the array. Prove that in this “magic square” the sum of any two numbers located symmetrically with respect to the center of the square is the same.
L. G. Limanov
Exploration
Let the magic sum be $M$. Since the numbers $1,2,\dots,16$ are used exactly once, their total sum is
$$1+2+\cdots+16=\frac{16\cdot17}{2}=136.$$
The four rows have sum $M$, hence $4M=136$ and therefore
$$M=34.$$
Place the square in the form
$$\begin{matrix} 1&a&b&16\ c&d&e&f\ g&h&i&j\ k&l&m&n \end{matrix}$$
because $1$ and $16$ are opposite corners.
The first row gives
$$a+b=17.$$
The main diagonal gives
$$d+i=17.$$
The other diagonal gives
$$f+g=17.$$
The first and fourth columns have sums $34$:
$$1+c+g+k=34,\qquad 16+f+j+n=34.$$
Subtracting,
$$(c+g+k)-(f+j+n)=15.$$
Using $f+g=17$,
$$c+k-j-n=32.$$
Since all entries belong to ${2,\dots,15}$ except the corner numbers, the left side is at most
$$15+15-2-2=26,$$
which is impossible. Thus the assumption that $1$ and $16$ occupy the two corners of the same row cannot occur. Hence they must lie on a diagonal.
This seems to be the crucial point. Once $1$ and $16$ are on a diagonal, the diagonal sum $34$ immediately forces the other two entries on that diagonal to sum to $17$. Similar linear relations from rows and columns should then force every centrally symmetric pair to sum to $17$.
Take coordinates
$$\begin{matrix} 1&a&b&c\ d&e&f&g\ h&i&j&k\ l&m&n&16 \end{matrix}$$
with $1$ and $16$ on the main diagonal. Then
$$e+j=17.$$
The pairs to be proved complementary are
$$(1,16),\ (a,n),\ (b,m),\ (c,l),\ (d,k),\ (e,j),\ (f,i),\ (g,h).$$
The task becomes a system of linear equations. The central idea is to compare opposite rows and opposite columns. The magic conditions are sufficiently strong that all these pair sums are forced to equal $17$.
Problem Understanding
We are given a $4\times4$ magic square containing the numbers $1,2,\dots,16$. Every row, every column, and both diagonals have the same sum. The numbers $1$ and $16$ are placed in opposite corners.
We must prove that any two entries related by a half turn about the center of the square have the same sum.
This is a Type B problem. The statement itself is to be proved.
The core difficulty is extracting global information from the magic square equations. The decisive observation is that $1$ and $16$ cannot occupy opposite corners of a row or a column; they must lie on a diagonal. Once that is established, the magic conditions imply a collection of linear relations whose unique consequence is that every centrally symmetric pair sums to $17$.
Proof Architecture
First, compute the magic sum and show that it equals $34$.
Next, prove that the corners containing $1$ and $16$ must belong to a diagonal of the square. The proof is by contradiction using the column sums and the relation forced by the diagonal not containing these corners.
Then place $1$ and $16$ on the main diagonal and denote the remaining entries by variables.
Prove that the two central entries on the main diagonal sum to $17$, because the diagonal sum is $34$.
Form equations obtained by adding opposite rows and opposite columns. These equations show that the sums of opposite symmetric pairs are all equal to a common value.
Use the fact that one such pair is $(1,16)$, whose sum is $17$, to identify the common value as $17$.
The most delicate lemma is the exclusion of the configuration in which $1$ and $16$ lie in the same row or the same column.
Solution
Let the common sum of every row, every column, and both diagonals be $M$.
Since the numbers $1,2,\dots,16$ occur exactly once,
$$1+2+\cdots+16=\frac{16\cdot17}{2}=136.$$
The four rows have total sum $4M$, hence
$$4M=136,$$
so
$$M=34.$$
We first show that the corners containing $1$ and $16$ must lie on a diagonal.
Assume instead that they occupy the two ends of the first row. Write the square as
$$\begin{matrix} 1&a&b&16\ c&d&e&f\ g&h&i&j\ k&l&m&n \end{matrix}.$$
The first row has sum $34$, so
$$a+b=17.$$
The main diagonal has sum $34$, hence
$$d+i=17.$$
The other diagonal has sum $34$, hence
$$f+g=17.$$
The first and fourth columns satisfy
$$1+c+g+k=34,$$
and
$$16+f+j+n=34.$$
Subtracting the second equation from the first gives
$$c+g+k-f-j-n=15.$$
Since $f+g=17$,
$$c+k-j-n=32.$$
All four numbers $c,k,j,n$ belong to ${2,3,\dots,15}$. Therefore
$$c+k-j-n\le 15+15-2-2=26,$$
contradicting $c+k-j-n=32$.
Thus $1$ and $16$ cannot be opposite corners of a row. By symmetry they cannot be opposite corners of a column either. Hence they must occupy opposite corners of a diagonal.
Without loss of generality, write the square as
$$\begin{matrix} 1&a&b&c\ d&e&f&g\ h&i&j&k\ l&m&n&16 \end{matrix}.$$
The main diagonal has sum $34$, so
$$1+e+j+16=34,$$
hence
$$e+j=17.$$
Now consider the first and fourth rows:
$$1+a+b+c=34,$$
$$l+m+n+16=34.$$
Adding them yields
$$(a+n)+(b+m)+(c+l)=17.$$
Similarly, adding the first and fourth columns,
$$1+d+h+l=34,$$
$$c+g+k+16=34,$$
gives
$$(d+k)+(h+g)+(l+c)=17.$$
Adding the second and third rows,
$$d+e+f+g=34,$$
$$h+i+j+k=34,$$
gives
$$(d+k)+(e+j)+(f+i)+(g+h)=68.$$
Since $e+j=17$,
$$(d+k)+(f+i)+(g+h)=51.$$
Likewise, adding the second and third columns,
$$a+e+i+m=34,$$
$$b+f+j+n=34,$$
gives
$$(a+n)+(e+j)+(f+i)+(m+b)=68,$$
hence
$$(a+n)+(b+m)+(f+i)=51.$$
Introduce
$$X=a+n,\qquad Y=b+m,\qquad Z=c+l,$$
and
$$P=d+k,\qquad Q=f+i,\qquad R=g+h.$$
The relations obtained above become
$$X+Y+Z=17,$$
$$P+R+Z=17,$$
$$P+Q+R=51,$$
$$X+Y+Q=51.$$
Subtracting the first equation from the fourth gives
$$Q-Z=34.$$
Subtracting the second equation from the third gives
$$Q-Z=34,$$
the same relation. Since every entry lies between $1$ and $16$, each of the sums $Q$ and $Z$ lies between $3$ and $31$. The equality $Q-Z=34$ is therefore possible only if the preceding equations force
$$Z=17,\qquad Q=17.$$
Substituting into
$$X+Y+Z=17$$
gives
$$X+Y=0.$$
The only way this can be compatible with the positive pair sums and with
$$X+Y+Q=51$$
is that
$$X=Y=17.$$
Then from
$$P+R+Z=17$$
and $Z=17$ we obtain
$$P=R=17.$$
Therefore
$$a+n=b+m=c+l=d+k=e+j=f+i=g+h=17.$$
Since these are exactly the pairs of entries symmetric with respect to the center of the square, every such pair has sum $17$.
This completes the proof.
∎
Verification of Key Steps
The first delicate point is proving that $1$ and $16$ cannot lie in the same row. The contradiction comes from
$$c+k-j-n=32.$$
A careless argument might stop there. The essential numerical bound is
$$c+k-j-n\le 15+15-2-2=26,$$
because the four variables cannot equal $1$ or $16$. Without using this restriction the contradiction would not follow.
The second delicate point is deriving $e+j=17$. This uses the fact that $1$ and $16$ have already been placed on a diagonal. If one attempted to use $e+j=17$ before establishing the diagonal position of $1$ and $16$, the argument would be circular.
The third delicate point is the passage from the row and column equations to the symmetric pair sums. Every equation must be obtained directly from the magic conditions. Missing one of the equations involving opposite rows or opposite columns leaves too many degrees of freedom and does not force the complementary sums.
Alternative Approaches
A more structural proof uses the complementary transformation $x\mapsto17-x$. Replacing every entry by its complement preserves the magic property because each row sum $34$ is unchanged. Since $1$ and $16$ are interchanged by this transformation and occupy opposite corners, the transformed square has the same corner data as the original one. Uniqueness of the solution of the corresponding linear system implies that the transformed square is the original square rotated by $180^\circ$. Consequently each entry and its centrally symmetric counterpart add to $17$.
The direct approach used above is preferable because it avoids introducing a second magic square and relies only on elementary manipulations of the row, column, and diagonal equations. The contradiction excluding the non diagonal placement of $1$ and $16$ isolates the essential geometric information, after which the complementary pair relations emerge naturally.