6
$\begingroup$

Here is a truly minimalistic and seemingly basic question which should have a simple solution (I hope it does).

Let $A$ be a finite set of integers with the smallest element $0$ and the largest element $l$. The sumset $C:=3A$ resides in the interval $[0,3l]$, and we let $C_1:=C\cap[0,l]$, $C_2:=C\cap[l,2l]$, and $C_3:=C\cap[2l,3l]$.

Is it true that $|C_2|\ge\frac12\,(|C_1|+|C_3|)$, for any choice of the set $A$?

Computations seem to suggest that the answer is in the affirmative.

$\endgroup$

1 Answer 1

6
$\begingroup$

No. Take $A = \{0,1,\ldots,9,10,20,30,\ldots,90,100,200,300,\ldots,900,1000\}$. Then $|C_1|=1001$, $|C_2|=272$ and $|C_3|=29$.

A smaller counterexample in the same spirit is $\{0,1,2,3,4,5,10,15,20,25,50,75,100\}$, with sizes $101, 53, 13$.

$\endgroup$
5
  • $\begingroup$ Great! Do you have a counterexample for $|C_2|\ge\min\{|C_1|,|C_3|\}$? $\endgroup$
    – Seva
    May 29, 2021 at 17:41
  • 1
    $\begingroup$ Not at the moment, no. For all I know your milder inequality might be true. $\endgroup$ May 29, 2021 at 18:06
  • 3
    $\begingroup$ Just for the record, a counterexample to the "weak" version: $A=\{0, 1, 3, 4, 12, 15, 18, 19, 20\}$. Here $|C_1|=|C_3|=18$, but $|C_2|=17$. $\endgroup$
    – Seva
    May 30, 2021 at 6:27
  • 1
    $\begingroup$ Mysteriously, I count sizes 21, 20, 21 from your counterexample, but yes, it is still a counterexample to the weak version. $\endgroup$ May 30, 2021 at 9:03
  • 1
    $\begingroup$ You are right, 21/20/21. Interestingly, all examples that I found satisfy $|C_2|=|C_1|-1=|C_3|-1$. I wonder whether the weak inequality holds if we redefine $C_1:=C\cap[0,l-1]$ and $C_3:=C\cap[2l+1,3l]$. $\endgroup$
    – Seva
    May 30, 2021 at 9:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.