Art of Problem Solving

2025 AMC 12A Problems/Problem 25: Difference between revisions

Tiankaizhang (talk | contribs)
Undo revision 266767 by Tiankaizhang (talk)
Tag: Undo
J314andrews (talk | contribs)
 
(96 intermediate revisions by 12 users not shown)
Line 1: Line 1:
Polynomials <imath>P(x)</imath> and <imath>Q(x)</imath> each have degree <imath>3</imath> and leading coefficient <imath>1</imath>, and their roots are all elements of <imath>\{1,2,3,4,5\}</imath>. The function <imath>f(x) = \tfrac{P(x)}{Q(x)}</imath> has the property that there exist real numbers <imath>a < b < c < d</imath> such that the set of all real numbers <imath>x</imath> such that <imath>f(x) \leq 0</imath> consists of the closed interval <imath>[a,b]</imath> together with the open interval <imath>(c,d)</imath>. How many functions <imath>f(x)</imath> are possible?
== Problem ==


==Solution 1 by Victor Zhang ==
Polynomials <imath>P(x)</imath> and <imath>Q(x)</imath> each have degree <imath>3</imath> and leading coefficient <imath>1</imath>, and their roots are all elements of <imath>\{1,2,3,4,5\}</imath>. The function <imath>f(x) = \tfrac{P(x)}{Q(x)}</imath> has the property that there exist real numbers <imath>a < b < c < d</imath> such that the set of all real numbers <imath>x</imath> such that <imath>f(x) \leq 0</imath> consists of the closed interval <imath>[a,b]</imath> together with the open interval <imath>(c,d)</imath>. How many ordered pairs of polynomials <imath>(P, Q)</imath> are possible?
\documentclass[12pt]{article}
\usepackage{amsmath, amssymb, amsthm}
\usepackage{enumitem}
\begin{document}
\section*{Solution}


\subsection*{Step 1: Sign chart analysis}
<imath>\textbf{(A)}~7 \qquad \textbf{(B)}~9 \qquad \textbf{(C)}~11 \qquad \textbf{(D)}~12 \qquad \textbf{(E)}~13 \qquad \textbf{(F)}~8</imath>
From the structure <imath>f(x) \le 0</imath> on <imath>[a,b] \cup (c,d)</imath> and <imath>f(x) &gt; 0</imath> elsewhere,
we deduce:
\begin{itemize}
\item <imath>a</imath> and <imath>b</imath> are zeros of <imath>f</imath> (since we transition from <imath>f &gt; 0</imath> to <imath>f \le
0</imath> at <imath>a</imath> and from <imath>f \le 0</imath> to <imath>f &gt; 0</imath> at <imath>b</imath>).
\item <imath>c</imath> and <imath>d</imath> are poles or holes of <imath>f</imath> (since on <imath>(c,d)</imath> we have <imath>f \le
0</imath>, at <imath>c^+</imath> and <imath>d^-</imath> the sign is negative, and immediately outside the
interval <imath>f &gt; 0</imath>).
\end{itemize}
Thus the sign pattern is:\\
<imath>\textstyle
\begin{array}{cccccccccc}
&amp; (-\infty, a) &amp; a &amp; (a,b) &amp; b &amp; (b,c) &amp; c &amp; (c,d) &amp; d &amp; (d,\infty) \\
f(x) &amp; + &amp; 0 &amp; - &amp; 0 &amp; + &amp; \text{pole} &amp; - &amp; \text{pole} &amp; +
\end{array}
</imath>
\section*{Step 2: Structure of \( f(x) \)}
According to the given conditions, \( f(x) \) can be expressed as:
\[
f(x) = \frac{(x-p_1)(x-p_2)(x-p_3)}{(x-q_1)(x-q_2)(x-q_3)}.
\]
Combining the sign analysis from Step 1, we can rewrite this expression as:
\[
f(x) = \frac{(x-a)(x-b)(x-p_3)}{(x-c)(x-d)(x-q_3)}.
\]
Furthermore, we can break down the expression into two parts:
\[
f(x) = \frac{(x-a)(x-b)}{(x-c)(x-d)} \cdot \frac{x-p_3}{x-q_3}.
\]
Defining \( f(x) = f_1(x) \cdot f_2(x) \), we have:
\[
f_1(x) = \frac{(x-a)(x-b)}{(x-c)(x-d)}, \quad f_2(x) = \frac{x-p_3}{x-q_3}.
\]


\section*{Step 3: Analysis of \( p_3 \) and \( q_3 \)}
==Why MAA initially got this wrong==
\begin{itemize}
If you look at solution 2, they overcounted the second case, essentially saying that <imath>f</imath> had <imath>2</imath> choices because it could equal <imath>c</imath> or <imath>d</imath>, so the intended answer was <imath>3 + 5 \cdot 2 = 13.</imath> This would be correct if the problem was asking for the number of pairs of polynomials. But there is actually only <imath>1</imath> choice for <imath>f</imath> because no matter what variable we choose <imath>f</imath> to equal, both polynomial combinations result in the same function. So the correct answer was <imath>3 + 5 \cdot 1 = 8</imath> functions.
\item By analyzing the sign changes of \( f_1(x) \), we find that it completely
satisfies the sign chart for \( f(x) \). We can therefore conclude that \( f_2(x) \)
must be positive on every interval; otherwise, it would introduce an extra sign
change. This forces the condition \( p_3 = q_3 \), because if \( p_3 \neq q_3 \),
the factor \( f_2(x) \) would be negative on the interval between \( p_3 \) and \(
q_3 \). Let \( p_3 = q_3 = t \).
\item Furthermore, \( t \) cannot lie within \( [a,b] \cup (c,d) \). If it did, it
would contradict the given condition that the set \( \{x : f(x) \le 0\} \) consists of
the closed interval \( [a, b] \) and the open interval \( (c, d) \).
\item Additionally, \( t \) cannot be equal to \( a \) or \( b \). If it were, that
point would be a hole (a point of discontinuity) in the graph of \( f(x) \), which
contradicts the fact that \( [a, b] \) is a \textbf{closed} interval where \( f(x) \le 0
\). On a closed interval, the function must be defined at the endpoints.
\end{itemize}
In conclusion, \( t \) can only be equal to \( c \) or \( d \), or lie in an interval
outside of \( [a,b] \cup (c,d) \).
\section*{Step 4: The Possible Values of <imath>a</imath>, <imath>b</imath>, <imath>c</imath>, <imath>d</imath>, and <imath>t</imath>}
\begin{itemize}
\item Given <imath>a &lt; b &lt; c &lt; d</imath> and the condition that <imath>\{x: f(x) \leq 0\} = [a,b]
\cup (c,d)</imath>, the number of ways to choose <imath>a,b,c,d</imath> from <imath>\{1,2,3,4,5\}</imath> is:
\[
\binom{5}{4} = 5
\]
The five cases are:
<imath>[1,2]\cup (3,4)</imath>; <imath>[1,2]\cup (3,5)</imath>; <imath>[1,2]\cup (4,5)</imath>; <imath>[1,3]\cup (4,5)</imath>;
<imath>[2,3]\cup (4,5)</imath>
\item \textbf{Case 1:} <imath>[1,2]\cup (3,4)</imath>, <imath>t</imath> can be 3, 4, or 5.\\
This gives us 3 combinations:
\[
f(x) = \frac{(x-1)(x-2)(x-3)}{(x-3)(x-4)(x-3)}
\]
\[
f(x) = \frac{(x-1)(x-2)(x-4)}{(x-3)(x-4)(x-4)}
\]
\[
f(x) = \frac{(x-1)(x-2)(x-5)}{(x-3)(x-4)(x-5)}
\]
\item \textbf{Case 2:} <imath>[1,2]\cup (3,5)</imath>, <imath>t</imath> can be 3 or 5.\\


This gives us 2 combinations:
Desmos graph showing how these functions are exactly the same: https://www.desmos.com/calculator/w6sk67ly1i
\[
f(x) = \frac{(x-1)(x-2)(x-3)}{(x-3)(x-5)(x-3)}
\]
\[
f(x) = \frac{(x-1)(x-2)(x-5)}{(x-3)(x-5)(x-5)}
\]
\item \textbf{Case 3:} <imath>[1,2]\cup (4,5)</imath>, <imath>t</imath> can be 3, 4, or 5.\\
This gives us 3 combinations:
\[
f(x) = \frac{(x-1)(x-2)(x-3)}{(x-4)(x-5)(x-3)}
\]
\[
f(x) = \frac{(x-1)(x-2)(x-4)}{(x-4)(x-5)(x-4)}
\]
\[
f(x) = \frac{(x-1)(x-2)(x-5)}{(x-4)(x-5)(x-5)}
\]
\item \textbf{Case 4:} <imath>[1,3]\cup (4,5)</imath>, <imath>t</imath> can be 4 or 5.\\
This gives us 2 combinations:
\[
f(x) = \frac{(x-1)(x-3)(x-4)}{(x-4)(x-5)(x-4)}
\]
\[
f(x) = \frac{(x-1)(x-3)(x-5)}{(x-4)(x-5)(x-5)}
\]
\item \textbf{Case 5:} <imath>[2,3]\cup (4,5)</imath>, <imath>t</imath> can be 1, 4, or 5.\\
This gives us 3 combinations:
\[
f(x) = \frac{(x-2)(x-3)(x-1)}{(x-4)(x-5)(x-1)}
\]
\[
f(x) = \frac{(x-2)(x-3)(x-4)}{(x-4)(x-5)(x-4)}
\]
\[
f(x) = \frac{(x-2)(x-3)(x-5)}{(x-4)(x-5)(x-5)}
\]
\end{itemize}
\section*{Step 5: Final Answer}
Summing up all possible cases from Step 4 gives us <imath>3 + 2 + 3 + 2 + 3 = 13</imath>
possible functions <imath>f(x)</imath>.
Thus the answer is <imath>\boxed{E}</imath>.
\end{document}


\section*{Step 4: The possible of a,b,c,d, and t }
~[[User:grogg007|grogg007]]
\begin{itemize}
\item According to given <imath>a &lt; b &lt; c &lt; d </imath> and \( [a,b] \cup (c,d) \), the ways
of choose a,b,c,d from {1,2,3,4,5}:
<cmath> \binom{5}{4}=5</cmath>
List 5 cases as:\\$[1,2]\cup (3,4); [1,2]\cup (3,5); [1,2]\cup (4,5); [1,3]\cup
(4,5); [2,3]\cup (4,5)<imath>


\item Case 1: </imath>[1,2]\cup (3,4)<imath>, t can be 3,4, or 5.\\
==Solution 1 (Got Wrong)==
Give us 3 combnation:
None of the answer choices on the official test (which asked for the number of possible functions <imath>f(x)</imath>) were correct, but choice E would be correct if this problem asked for the number of pairs of functions <imath>(P(x), Q(x))</imath>.
\[
f(x) = \frac{(x-1)(x-2)(x-3)}{(x-3)(x-4)(x-3)}
\]
\[
f(x) = \frac{(x-1)(x-2)(x-4)}{(x-3)(x-4)(x-4
)}
\]
\[
f(x) = \frac{(x-1)(x-2)(x-5)}{(x-3)(x-4)(x-5)}.
\]
\item Case 2: </imath>[1,2]\cup (3,5)<imath>, t can be 3, or 5.\\
Give us 2 combnation:
\[
f(x) = \frac{(x-1)(x-2)(x-3)}{(x-3)(x-5)(x-3)}
\]
\[
f(x) = \frac{(x-1)(x-2)(x-5)}{(x-3)(x-4)(x-5
)}
\]
\item Case 3: </imath>[1,2]\cup (4,5)<imath>, t can be 3,4, or 5.\\
Give us 3 combnation:
\[
f(x) = \frac{(x-1)(x-2)(x-3)}{(x-4)(x-5)(x-3)}
\]
\[
f(x) = \frac{(x-1)(x-2)(x-4)}{(x-4)(x-5)(x-4
)}
\]
\[
f(x) = \frac{(x-1)(x-2)(x-5)}{(x-4)(x-5)(x-5)}.
\]
\item Case 4: </imath>[1,3]\cup (4,5)<imath>, t can be 4, or 5.\\
Give us 2 combnation:
\[
f(x) = \frac{(x-1)(x-3)(x-4)}{(x-4)(x-5)(x-4)}


\]
Let <imath>R(x) = \frac{P(x)}{Q(x)}</imath>.  Since <imath>R(x) \leq 0</imath> on <imath>[a, b]</imath> but not for values slightly less than <imath>a</imath> or slightly more than <imath>b</imath>, <imath>P(x) = 0</imath> at <imath>x = a</imath> and <imath>x = b</imath>.  Therefore, <imath>P(x) = (x-a)(x-b)(x-r)</imath> for some <imath>r \in \{1, 2, 3, 4, 5\}</imath>.
\[
f(x) = \frac{(x-1)(x-3)(x-5)}{(x-4)(x-5)(x-5
)}
\]
\item Case 5: </imath>[2,3]\cup (4,5)<imath>, t can be 1,4, or 5.\\
Give us 3 combnation:
\[
f(x) = \frac{(x-2)(x-3)(x-1)}{(x-4)(x-5)(x-1)}
\]
\[
f(x) = \frac{(x-2)(x-3)(x-4)}{(x-4)(x-5)(x-4)}
\] \[
f(x) = \frac{(x-2)(x-3)(x-5)}{(x-4)(x-5)(x-5)}
\]


\section*{Step 5: final answer }
Since <imath>R(x) \leq 0</imath> on <imath>(c, d)</imath> but not at <imath>x = c</imath> or <imath>x = d</imath>, <imath>R(x)</imath> is not continuous at <imath>x = c</imath> or <imath>x = d</imath>.  Therefore, <imath>R(x)</imath> must be undefined at <imath>x = c</imath> and <imath>x = d</imath>, that is, <imath>Q(x) = 0</imath> at <imath>x = c</imath> and <imath>x = d</imath>.  So <imath>Q(x) = (x-c)(x-d)(x-s)</imath> for some <imath>s \in \{1, 2, 3, 4, 5\}</imath>.
\Sum up all possible cases of step 4 give us 13 possible functions f(x).
 
Therefore, <imath>R(x) = \frac{(x-a)(x-b)(x-r)}{(x-c)(x-d)(x-s)}</imath>.  Notice that <imath>\frac{(x-a)(x-b)}{(x-c)(x-d)} \leq 0</imath> only on <imath>[a, b]</imath> and <imath>(c, d)</imath>.  Therefore, <imath>\frac{x-r}{x-s}</imath> must be nonnegative for all <imath>x \notin \{a, b, c, d, r, s\}</imath>.  This only happens if <imath>r = s</imath>.
 
Thus <imath>R(x) = \frac{(x-a)(x-b)(x-r)}{(x-c)(x-d)(x-r)}</imath>, which is the same as <imath>\frac{(x-a)(x-b)}{(x-c)(x-d)}</imath> except that it is undefined at <imath>x = r</imath>.  Thus <imath>R(x)</imath> satisfies the desired property as long as <imath>r \notin [a, b] \cup (c, d)</imath>.
 
Note that each quintuple <imath>(a, b, c, d, r)</imath> defines a unique pair of functions <imath>(P(x), Q(x))</imath>.
 
If <imath>(a, b, c, d) = (1, 2, 3, 4)</imath>, <imath>r</imath> can be <imath>3</imath>, <imath>4</imath>, or <imath>5</imath>.
 
If <imath>(a, b, c, d) = (1, 2, 3, 5)</imath>, <imath>r</imath> can be <imath>3</imath> or <imath>5</imath>.
 
If <imath>(a, b, c, d) = (1, 2, 4, 5)</imath>, <imath>r</imath> can be <imath>3</imath>, <imath>4</imath>, or <imath>5</imath>.
 
If <imath>(a, b, c, d) = (1, 3, 4, 5)</imath>, <imath>r</imath> can be <imath>4</imath> or <imath>5</imath>.
 
If <imath>(a, b, c, d) = (2, 3, 4, 5)</imath>, <imath>r</imath> can be <imath>1</imath>, <imath>4</imath>, or <imath>5</imath>.
 
Therefore, there are <imath>\boxed{(\textbf{E})\ 13}</imath> possible pairs of functions <imath>(P(x), Q(x))</imath>.
 
-j314andrews
 
==Solution 2 (credit to Sohil Rathi's video solution, pi_is_3.14 / OmegaLearn)==
This is the solution to the original problem, which asked for the number of possible '''functions''', not pairs of polynomials.
 
First, consider the problem if it were talking about two second degree polynomials. We can see that the function <cmath>f(x) = \frac{(x-a)(x-b)}{(x-c)(x-d)}</cmath> by itself satisfies the condition of dropping below the x axis over <imath>[a,b] \cup (c,d).</imath> Now, we need to add one extra <imath>(x-n)</imath> term each to the numerator and denominator.
 
Case 1:
<imath>f(x) = \frac{(x-a)(x-b)(x-f)}{(x-c)(x-d)(x-f)}</imath> for some other value <imath>f</imath> not equal to <imath>a, b, c,</imath> or <imath>d.</imath>
 
Note that <imath>f</imath> cannot be between <imath>a</imath> and <imath>b</imath> or between <imath>c</imath> and <imath>d</imath> because this would create a hole in the interval. So the only possibilities are:
 
*<imath>a = 1, b = 2, c = 3, d = 4, f = 5</imath>
*<imath>a = 1, b = 2, c = 4, d = 5, f = 3</imath>
*<imath>a = 2, b = 3, c = 4, d = 5, f = 1</imath>
 
This gives us <imath>3</imath> valid functions so far.
 
 
Case 2:
<imath>f(x) = \frac{(x-a)(x-b)(x-f)}{(x-c)(x-d)(x-f)}</imath> where <imath>f</imath> is <imath>c</imath> or <imath>d.</imath>
 
If <imath>f</imath> equals <imath>c</imath> or <imath>d</imath> the resulting function is essentially equivalent to <imath>f(x) = \frac{(x-a)(x-b)}{(x-c)(x-d)}</imath> because a hole would already exist, so we can safely cancel the <imath>(x-f)</imath> terms. It doesn't matter whether we choose <imath>f</imath> to equal <imath>c</imath> or <imath>d</imath> because the function will still be the same. There are <imath>\binom{5}{4} = 5</imath> ways to select the values for <imath>a, b, c,</imath> and <imath>d</imath> and <imath>1</imath> way to choose any variable for <imath>f</imath> to equal, giving <imath>5 \cdot 1 = 5</imath> valid functions for this case.
 
 
The correct answer is <imath>5 + 3 = \boxed{\textbf{(F) } 8}</imath> functions.
 
~[[User:grogg007|grogg007]], original solution by OmegaLearn/ Sohil Rathi
 
i tried my best to show what I think was Sohil's thought process, but please feel free to edit this if you can explain it better :)
 
==Solution 3==
 
'''Solution 1''' by user j314andrews is incomplete; '''Solution 2''' by user [[User:grogg007|grogg007]] commits a serious logical error because it only provides functions that satisfy the requirements, without showing that no other functions do. Other places like [https://wiki.randommath.com/amc12/2025/part-a/problem-25] gives a good solution, but unfortunately it requires calculus. Here I am trying to provide a good solution that does not require calculus.
 
 
Let <imath>D</imath> and <imath>E</imath> be the set of the real roots of <imath>P(x)</imath> and <imath>Q(x)</imath>, respectively. Then, outside of the intervals <imath>[a,b]</imath> and <imath>(c,d)</imath>, <imath>f(x)>0</imath> except when <imath>x \in E</imath>, where <imath>f(x)</imath> is undefined.
 
 
For any function <imath>g(x)</imath>, define a real number <imath>t</imath> as a '''toggling point''' for <imath>g(x)</imath> if there exists a sequence <imath>t_1 < t_2 < ... < t_N</imath>, with the understanding that <imath>t_0 = -\infty</imath> and <imath>t_{N+1} = \infty</imath>, such that
* <imath>t = t_k</imath> for some <imath>1 \leq k \leq N</imath>,
* <imath>g(x)</imath> is of the same sign (either always positive or always negative) on each interval <imath>(t_j, t_{j+1})</imath>, for <imath>j=0, ..., N</imath>, and
* <imath>g(x)</imath> has opposite signs on <imath>(t_{k-1}, t_k)</imath> and <imath>(t_k, t_{k+1})</imath>.
 
We call these open intervals <imath>(t_j, t_{j+1})</imath> (<imath>j=0, ..., N</imath>) the '''bands''' of <imath>g(x)</imath>.
 
 
The polynomial <imath>P(x)</imath> has no toggling point outside of <imath>D</imath>, and <imath>Q(x)</imath> has no toggling point outside of <imath>E</imath>. Now in our problem, <imath>f(x)</imath> has exactly 4 toggling points: <imath>a, b, c, d</imath>. Our task is to see how these four points are related to the sets <imath>D</imath> and <imath>E</imath>. 
 
'''Lemma'''. At each toggling point <imath>t</imath> for <imath>f(x)</imath>,
# either <imath>P(t)=0</imath> or <imath>Q(t)=0</imath> (or both);
# if the multiplicities of the root <imath>t</imath> of <imath>P(x)</imath> and <imath>Q(x)</imath> are <imath>m</imath> and <imath>n</imath>, respectively, then <imath>m-n</imath> is odd.
 
'''Proof'''. (1) if both <imath>P(t)</imath> and <imath>Q(t)</imath> are nonzero, then <imath>t</imath> belongs to a band of <imath>P(x)</imath> and a band of <imath>Q(x)</imath>. Taking the intersection between these two bands, we get an open interval on which <imath>f(x)</imath> has the same sign as <imath>f(t)</imath>. Thus <imath>t</imath> cannot be a toggling point of <imath>f(x)</imath>.
 
(2) If <imath>m</imath> is odd, there are real numbers <imath>r, s</imath>, with <imath>r<t<s</imath>, such that <imath>P(x)</imath> has the same sign on <imath>(r,t)</imath> and has the opposite sign on <imath>(t, s)</imath>. If <imath>m</imath> is even, there are real numbers <imath>r, s</imath>, with <imath>r<t<s</imath>, such that <imath>P(x)</imath> has the same sign on <imath>(r,t)</imath> and on <imath>(t,s)</imath>. Similar statements can be made of <imath>Q(x)</imath>, having intervals <imath>(u, t)</imath> and <imath>(t, v)</imath> so that <imath>Q(t)</imath> is of the same sign on <imath>(u,t)</imath>, and of either a different or the same sign on <imath>(t, v)</imath>, depending on the parity of <imath>n</imath>.  If both <imath>m</imath> and <imath>n</imath> are odd or both <imath>m</imath> and <imath>n</imath> are even, then <imath>P(t)/Q(t)</imath> is of the same sign on the interval <imath>(\max\{r, u\}, t)</imath> and <imath>(t, \min\{s, v\})</imath>, so <imath>t</imath> cannot be a toggling point for <imath>f(x)</imath>.
 
 
'''Solution to the problem'''.
Since <imath>a</imath> is a toggling point for <imath>f(x)</imath>, either <imath>P(a)=0</imath> or <imath>Q(a)=0</imath>, or both. But since <imath>f(a)</imath> is defined, we have <imath>Q(a)\neq 0</imath>, so <imath>P(a)=0</imath>. The same can be said of <imath>b</imath>, so <imath>P(b)=0</imath>. The multiplicities of <imath>a</imath> and <imath>b</imath> must be <imath>1</imath>. Otherwise, point (2) of the Lemma would require the multiplicity of one of them to be at least <imath>3</imath>, but <imath>P(x)</imath> is only cubic.
 
 
<imath>c</imath> is a toggling point for <imath>f(x)</imath>, so either <imath>P(c)=0</imath> or <imath>Q(c)=0</imath>, or both. But if <imath>Q(c)\neq 0</imath>, we have <imath>f(c)=0</imath>, violating the condition. So <imath>Q(c)=0</imath>. Similarly, <imath>Q(d)=0</imath>. Concerning their multiplicities, point (2) of the Lemma gives us three cases, given that <imath>Q(x)</imath> is only cubic:
* (a) <imath>c</imath> is a single root of <imath>Q(x)</imath>, <imath>P(c)\neq 0</imath>, <imath>d</imath> is a double root of <imath>Q(x)</imath> and a single root of <imath>P(x)</imath>;
* (b) switching the roles of <imath>c</imath> and <imath>d</imath>;
* (c) both <imath>c</imath> and <imath>d</imath> are single roots of <imath>Q(x)</imath>, and <imath>Q(x)</imath> has a single root <imath>e</imath> other than <imath>c</imath> and <imath>d</imath>.
 
 
Case (a) provides functions in the form of <imath>\frac{(x-a)(x-b)(x-d)}{(x-c)(x-d)^2}</imath>. There are 5 of them because there are 5 ways to choose <imath>a, b, c, d</imath> from the set <imath>\{1, 2, 3, 4, 5\}</imath>.
 
Case (b) provides functions in the form of <imath>\frac{(x-a)(x-b)(x-c)}{(x-c)^2(x-d)}</imath>, but they are the same as those provided in case (a). To ensure this, notice that the two expressions have the same domain and they are equal for every <imath>x</imath> in their domains.
 
Case (c) requires that <imath>P(e)=0</imath>, because if <imath>P(e)\neq 0</imath> then <imath>e</imath> becomes an extra toggling point for <imath>f(x)</imath>. So this case provides functions in the form of <imath>\frac{(x-a)(x-b)(x-e)}{(x-c)(x-d)(x-e)}</imath>. But since <imath>f(e)</imath> is undefined, we need to make sure that <imath>e</imath> does not fall into <imath>[a,b]</imath> or <imath>(c,d)</imath>. So <imath>e</imath> can be <imath>1, 3, 5</imath>. In all these cases, we also have <imath>f(e) = \frac{(e-a)(e-b)}{(e-c)(e-d)} > 0</imath>, satisfying the conditions. That's 3 more functions.
 
So the total number of functions is <imath>5 + 3 = \boxed{8}</imath>.
 
[[User:Lightest|Lightest]] ([[User talk:Lightest|talk]]) 22:51, 10 November 2025 (EST)


Thus the answer is </imath>\boxed{E}$.
\end{document}


==Video Solution 1 by OmegaLearn==
==Video Solution 1 by OmegaLearn==
https://youtu.be/SPbTyq3Dz_0
https://youtu.be/SPbTyq3Dz_0
==See Also==
{{AMC12 box|year=2025|ab=A|num-b=24|after=Last Problem}}
{{MAA Notice}}

Latest revision as of 12:14, 11 November 2025

Problem

Polynomials $P(x)$ and $Q(x)$ each have degree $3$ and leading coefficient $1$, and their roots are all elements of $\{1,2,3,4,5\}$. The function $f(x) = \tfrac{P(x)}{Q(x)}$ has the property that there exist real numbers $a < b < c < d$ such that the set of all real numbers $x$ such that $f(x) \leq 0$ consists of the closed interval $[a,b]$ together with the open interval $(c,d)$. How many ordered pairs of polynomials $(P, Q)$ are possible?

$\textbf{(A)}~7 \qquad \textbf{(B)}~9 \qquad \textbf{(C)}~11 \qquad \textbf{(D)}~12 \qquad \textbf{(E)}~13 \qquad \textbf{(F)}~8$

Why MAA initially got this wrong

If you look at solution 2, they overcounted the second case, essentially saying that $f$ had $2$ choices because it could equal $c$ or $d$, so the intended answer was $3 + 5 \cdot 2 = 13.$ This would be correct if the problem was asking for the number of pairs of polynomials. But there is actually only $1$ choice for $f$ because no matter what variable we choose $f$ to equal, both polynomial combinations result in the same function. So the correct answer was $3 + 5 \cdot 1 = 8$ functions.

Desmos graph showing how these functions are exactly the same: https://www.desmos.com/calculator/w6sk67ly1i

~grogg007

Solution 1 (Got Wrong)

None of the answer choices on the official test (which asked for the number of possible functions $f(x)$) were correct, but choice E would be correct if this problem asked for the number of pairs of functions $(P(x), Q(x))$.

Let $R(x) = \frac{P(x)}{Q(x)}$. Since $R(x) \leq 0$ on $[a, b]$ but not for values slightly less than $a$ or slightly more than $b$, $P(x) = 0$ at $x = a$ and $x = b$. Therefore, $P(x) = (x-a)(x-b)(x-r)$ for some $r \in \{1, 2, 3, 4, 5\}$.

Since $R(x) \leq 0$ on $(c, d)$ but not at $x = c$ or $x = d$, $R(x)$ is not continuous at $x = c$ or $x = d$. Therefore, $R(x)$ must be undefined at $x = c$ and $x = d$, that is, $Q(x) = 0$ at $x = c$ and $x = d$. So $Q(x) = (x-c)(x-d)(x-s)$ for some $s \in \{1, 2, 3, 4, 5\}$.

Therefore, $R(x) = \frac{(x-a)(x-b)(x-r)}{(x-c)(x-d)(x-s)}$. Notice that $\frac{(x-a)(x-b)}{(x-c)(x-d)} \leq 0$ only on $[a, b]$ and $(c, d)$. Therefore, $\frac{x-r}{x-s}$ must be nonnegative for all $x \notin \{a, b, c, d, r, s\}$. This only happens if $r = s$.

Thus $R(x) = \frac{(x-a)(x-b)(x-r)}{(x-c)(x-d)(x-r)}$, which is the same as $\frac{(x-a)(x-b)}{(x-c)(x-d)}$ except that it is undefined at $x = r$. Thus $R(x)$ satisfies the desired property as long as $r \notin [a, b] \cup (c, d)$.

Note that each quintuple $(a, b, c, d, r)$ defines a unique pair of functions $(P(x), Q(x))$.

If $(a, b, c, d) = (1, 2, 3, 4)$, $r$ can be $3$, $4$, or $5$.

If $(a, b, c, d) = (1, 2, 3, 5)$, $r$ can be $3$ or $5$.

If $(a, b, c, d) = (1, 2, 4, 5)$, $r$ can be $3$, $4$, or $5$.

If $(a, b, c, d) = (1, 3, 4, 5)$, $r$ can be $4$ or $5$.

If $(a, b, c, d) = (2, 3, 4, 5)$, $r$ can be $1$, $4$, or $5$.

Therefore, there are $\boxed{(\textbf{E})\ 13}$ possible pairs of functions $(P(x), Q(x))$.

-j314andrews

Solution 2 (credit to Sohil Rathi's video solution, pi_is_3.14 / OmegaLearn)

This is the solution to the original problem, which asked for the number of possible functions, not pairs of polynomials.

First, consider the problem if it were talking about two second degree polynomials. We can see that the function \[f(x) = \frac{(x-a)(x-b)}{(x-c)(x-d)}\] by itself satisfies the condition of dropping below the x axis over $[a,b] \cup (c,d).$ Now, we need to add one extra $(x-n)$ term each to the numerator and denominator.

Case 1: $f(x) = \frac{(x-a)(x-b)(x-f)}{(x-c)(x-d)(x-f)}$ for some other value $f$ not equal to $a, b, c,$ or $d.$

Note that $f$ cannot be between $a$ and $b$ or between $c$ and $d$ because this would create a hole in the interval. So the only possibilities are:

  • $a = 1, b = 2, c = 3, d = 4, f = 5$
  • $a = 1, b = 2, c = 4, d = 5, f = 3$
  • $a = 2, b = 3, c = 4, d = 5, f = 1$

This gives us $3$ valid functions so far.


Case 2: $f(x) = \frac{(x-a)(x-b)(x-f)}{(x-c)(x-d)(x-f)}$ where $f$ is $c$ or $d.$

If $f$ equals $c$ or $d$ the resulting function is essentially equivalent to $f(x) = \frac{(x-a)(x-b)}{(x-c)(x-d)}$ because a hole would already exist, so we can safely cancel the $(x-f)$ terms. It doesn't matter whether we choose $f$ to equal $c$ or $d$ because the function will still be the same. There are $\binom{5}{4} = 5$ ways to select the values for $a, b, c,$ and $d$ and $1$ way to choose any variable for $f$ to equal, giving $5 \cdot 1 = 5$ valid functions for this case.


The correct answer is $5 + 3 = \boxed{\textbf{(F) } 8}$ functions.

~grogg007, original solution by OmegaLearn/ Sohil Rathi

i tried my best to show what I think was Sohil's thought process, but please feel free to edit this if you can explain it better :)

Solution 3

Solution 1 by user j314andrews is incomplete; Solution 2 by user grogg007 commits a serious logical error because it only provides functions that satisfy the requirements, without showing that no other functions do. Other places like [1] gives a good solution, but unfortunately it requires calculus. Here I am trying to provide a good solution that does not require calculus.


Let $D$ and $E$ be the set of the real roots of $P(x)$ and $Q(x)$, respectively. Then, outside of the intervals $[a,b]$ and $(c,d)$, $f(x)>0$ except when $x \in E$, where $f(x)$ is undefined.


For any function $g(x)$, define a real number $t$ as a toggling point for $g(x)$ if there exists a sequence $t_1 < t_2 < ... < t_N$, with the understanding that $t_0 = -\infty$ and $t_{N+1} = \infty$, such that

  • $t = t_k$ for some $1 \leq k \leq N$,
  • $g(x)$ is of the same sign (either always positive or always negative) on each interval $(t_j, t_{j+1})$, for $j=0, ..., N$, and
  • $g(x)$ has opposite signs on $(t_{k-1}, t_k)$ and $(t_k, t_{k+1})$.

We call these open intervals $(t_j, t_{j+1})$ ($j=0, ..., N$) the bands of $g(x)$.


The polynomial $P(x)$ has no toggling point outside of $D$, and $Q(x)$ has no toggling point outside of $E$. Now in our problem, $f(x)$ has exactly 4 toggling points: $a, b, c, d$. Our task is to see how these four points are related to the sets $D$ and $E$.

Lemma. At each toggling point $t$ for $f(x)$,

  1. either $P(t)=0$ or $Q(t)=0$ (or both);
  2. if the multiplicities of the root $t$ of $P(x)$ and $Q(x)$ are $m$ and $n$, respectively, then $m-n$ is odd.

Proof. (1) if both $P(t)$ and $Q(t)$ are nonzero, then $t$ belongs to a band of $P(x)$ and a band of $Q(x)$. Taking the intersection between these two bands, we get an open interval on which $f(x)$ has the same sign as $f(t)$. Thus $t$ cannot be a toggling point of $f(x)$.

(2) If $m$ is odd, there are real numbers $r, s$, with $r<t<s$, such that $P(x)$ has the same sign on $(r,t)$ and has the opposite sign on $(t, s)$. If $m$ is even, there are real numbers $r, s$, with $r<t<s$, such that $P(x)$ has the same sign on $(r,t)$ and on $(t,s)$. Similar statements can be made of $Q(x)$, having intervals $(u, t)$ and $(t, v)$ so that $Q(t)$ is of the same sign on $(u,t)$, and of either a different or the same sign on $(t, v)$, depending on the parity of $n$. If both $m$ and $n$ are odd or both $m$ and $n$ are even, then $P(t)/Q(t)$ is of the same sign on the interval $(\max\{r, u\}, t)$ and $(t, \min\{s, v\})$, so $t$ cannot be a toggling point for $f(x)$.


Solution to the problem. Since $a$ is a toggling point for $f(x)$, either $P(a)=0$ or $Q(a)=0$, or both. But since $f(a)$ is defined, we have $Q(a)\neq 0$, so $P(a)=0$. The same can be said of $b$, so $P(b)=0$. The multiplicities of $a$ and $b$ must be $1$. Otherwise, point (2) of the Lemma would require the multiplicity of one of them to be at least $3$, but $P(x)$ is only cubic.


$c$ is a toggling point for $f(x)$, so either $P(c)=0$ or $Q(c)=0$, or both. But if $Q(c)\neq 0$, we have $f(c)=0$, violating the condition. So $Q(c)=0$. Similarly, $Q(d)=0$. Concerning their multiplicities, point (2) of the Lemma gives us three cases, given that $Q(x)$ is only cubic:

  • (a) $c$ is a single root of $Q(x)$, $P(c)\neq 0$, $d$ is a double root of $Q(x)$ and a single root of $P(x)$;
  • (b) switching the roles of $c$ and $d$;
  • (c) both $c$ and $d$ are single roots of $Q(x)$, and $Q(x)$ has a single root $e$ other than $c$ and $d$.


Case (a) provides functions in the form of $\frac{(x-a)(x-b)(x-d)}{(x-c)(x-d)^2}$. There are 5 of them because there are 5 ways to choose $a, b, c, d$ from the set $\{1, 2, 3, 4, 5\}$.

Case (b) provides functions in the form of $\frac{(x-a)(x-b)(x-c)}{(x-c)^2(x-d)}$, but they are the same as those provided in case (a). To ensure this, notice that the two expressions have the same domain and they are equal for every $x$ in their domains.

Case (c) requires that $P(e)=0$, because if $P(e)\neq 0$ then $e$ becomes an extra toggling point for $f(x)$. So this case provides functions in the form of $\frac{(x-a)(x-b)(x-e)}{(x-c)(x-d)(x-e)}$. But since $f(e)$ is undefined, we need to make sure that $e$ does not fall into $[a,b]$ or $(c,d)$. So $e$ can be $1, 3, 5$. In all these cases, we also have $f(e) = \frac{(e-a)(e-b)}{(e-c)(e-d)} > 0$, satisfying the conditions. That's 3 more functions.

So the total number of functions is $5 + 3 = \boxed{8}$.

Lightest (talk) 22:51, 10 November 2025 (EST)


Video Solution 1 by OmegaLearn

https://youtu.be/SPbTyq3Dz_0

See Also

2025 AMC 12A (ProblemsAnswer KeyResources)
Preceded by
Problem 24
Followed by
Last Problem
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
All AMC 12 Problems and Solutions

These problems are copyrighted © by the Mathematical Association of America.