The players are given

$ ax^3 + bx^2+ cx+ d.$

Players alternate taking turns putting in a real coefficient, including zero.

Player 2 will win if the resulting formula has exactly one real zero, else player 1 will win.

I know at least player 1 cannot let player 2 have the constant last, his first move also cannot be $a=0$ and that if player 1 takes $a \ne 0$ and player 2 sets $d=0$ player 1 can win by $sign(c)=-sign(a)$ . However I feel like I'm looking at too many options and not zeroing at the right solution.

  • $\begingroup$ resulting formula has atleast one zero or exactly one zero? $\endgroup$ Oct 21 '17 at 10:01
  • $\begingroup$ Exactly one real zero. $\endgroup$ Oct 21 '17 at 10:01
  • $\begingroup$ Can you explain the rules of the game ? $\endgroup$
    – user65203
    Oct 21 '17 at 10:02
  • $\begingroup$ Given the polynomial in the original post, we take turns given a value to the coefficients. Such as Player 1 chooses $a=1$ Player 2 chooses $d=0$ Player 1 chooses $c=-1$ No matter what player 2 chooses the polynomial will have two zeroes. Therefore player 1 wins. However if player two manages to have just one zero in the resulting polynomial he will win. $\endgroup$ Oct 21 '17 at 10:08
  • 3
    $\begingroup$ Is a double root counted as one or two ? $\endgroup$
    – user65203
    Oct 21 '17 at 10:25

Player 2 has a winning tactic. This player can namely force a polynomial of the form $ax^3+bx^2+ax+b = (x^2+1)(ax+b)$ with $a \neq 0$, $bx^2+2bx+b = b(x+1)^2$ with $b \neq 0$, $2bx^3+bx^2+b = b(2x^2-x+1)(x+1)$ with $b \neq 0$, or $e\cdot x^i$ with $i > 0$. All these polynomials have exactly one real zero. This can be done with the following tactic:

Suppose that player 1 starts with setting one coefficient to $0$. Now we apply the following tactic:

First move:

  • If player 1 set $d=0$, player 2 can take $a=0$.

  • If player set some other coefficient to $0$, player 2 takes $d=0$.

Second move:

  • If player 1 set some other coefficient to $0$, player 2 sets the remaining coefficient to $1$, resulting in a polynomial $x^i$ for some $i >0$ (because we forced $d = 0$ in the first move).

  • If player 1 does something else, then player 2 sets the remaining coefficient to $0$, which results in some polynomial $e\cdot x^i$ for some $i > 0$.

In the case that the first player did not start with a $0$, player 2 can do the following:

First move:

  • If player 1 picked $a$ or $c$, player 2 can take the other number of these two in such a way that $a = c$.

  • If player 1 picked $b$ or $d$, player 2 can also take the other number of these two to get $b = d$

Second move:

  • If in the first move we enforced that $a = c$, we can now easily take $b = d$. This results in a polynomial of the form $ax^3+bx^2+ax+b$ with $a \neq 0$.

  • Similarly, if we had $b = d$ after the first move, and player one did not take $a = 0$ or $c = 0$, we can now pick the remaining coefficient such that $a = c$, and again the result is of the form $ax^3+bx^2+ax+b$ with $a \neq 0$.

  • If the player 1 made the move $a = 0$, we set $c = 2b$, resulting in the polynomial $bx^2+2bx+b$ with $b \neq 0$.

  • Finally, if player 1 put $c = 0$, then we set $a = 2b$, giving the polynomial $2bx^3+bx^2+b$ with $b \neq 0$.

So we see that this tactic always results in a polynomial with exactly one zero.

  • 1
    $\begingroup$ one weak spot: if player 1 begins with $b = 1$, we choose $d = 1$. Then if player 1 chooses $0$ for either of the others, according to this strategy we should choose to make $a = c = 0$, but this gives a polynomial with no real roots. Easily fixable, though, I think. $\endgroup$ Oct 21 '17 at 12:20
  • $\begingroup$ @NickPavlov you are right, I will try to fix this. $\endgroup$
    – Reinier
    Oct 21 '17 at 12:54
  • $\begingroup$ I have now fixed the weak spot, thanks for spotting it! $\endgroup$
    – Reinier
    Oct 21 '17 at 13:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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