AMC 10 Step by Step

2011 AMC 10B Problem 24

P21-25~6 minPrint

A lattice point in an xyxy -coordinate system is any point (x,y)(x, y) where both xx and yy are integers. The graph of y=mx+2y = mx +2 passes through no lattice point with 0<x≤1000 < x \le 100 for all mm such that 1/2<m<a1/2 < m < a . What is the maximum possible value of aa ?

Log in to save your progress.
Show solution without answering

Key insight

The line hits a lattice point at x exactly when the slope times x is whole. So a is the smallest fraction above 1/2 with bottom at most 100.

Solution

Step A: when does the line hit a lattice point?

  1. For a whole number xx from 11 to 100100, the line goes through (x,mx+2)(x, mx + 2). This is a lattice point exactly when mx+2mx + 2 is a whole number, which means mxmx is a whole number.
  2. Example: if m=35m = \frac{3}{5}, then at x=5x = 5 we get mx=3mx = 3, and the line passes through (5,5)(5, 5).
  3. In general, mx=kmx = k (a whole number) means m=kxm = \frac{k}{x}. So the line hits a lattice point in the range exactly when mm can be written as a fraction with bottom number 100100 or less.

Step B: turn it into a fraction question.

  1. So we need every slope between 12\frac12 and aa to avoid all such fractions. The largest aa is the first such fraction above 12\frac12: the smallest fraction bigger than 12\frac12 with bottom at most 100100.

Step C: find the fraction closest to 12\frac12 from above.

  1. How far is kx\frac{k}{x} above 12\frac12? Write kx−12=2k−x2x\frac{k}{x} - \frac{1}{2} = \frac{2k - x}{2x}. The top, 2k−x2k - x, is a whole number at least 11.
  2. To make the gap small, we want the top as small as possible (2k−x=12k - x = 1) and the bottom as large as possible.
  3. 2k−x=12k - x = 1 means x=2k−1x = 2k - 1 is odd. The largest odd xx up to 100100 is 9999, with k=50k = 50. That gives 5099\frac{50}{99}, which is 1198\frac{1}{198} above 12\frac12.
  4. Could x=100x = 100 do better? Then 2k−1002k - 100 is even, so it is at least 22. The best is 51100\frac{51}{100}, which is 2200=1100\frac{2}{200} = \frac{1}{100} above 12\frac12. That is farther than 1198\frac{1}{198}.
  5. So 5099\frac{50}{99} is the smallest fraction above 12\frac12 with bottom at most 100100.

Step D: check both halves.

  1. Every mm between 12\frac12 and 5099\frac{50}{99} is safe, because there is no allowed fraction in between.
  2. If aa were bigger than 5099\frac{50}{99}, the slope m=5099m = \frac{50}{99} would be included, and that line passes through (99,50+2)=(99,52)(99, 50 + 2) = (99, 52). So aa cannot be bigger.

The answer is (B) 5099\boxed{\textbf{(B)}\ \frac{50}{99}}.

Why this works

A lattice-point question about a line with slope mm is really a question about which fractions mm can equal. The gap formula 2k−x2x\frac{2k - x}{2x} shows that the fraction just above 12\frac12 has top and bottom as close to "half" as possible, with the largest odd bottom number.

Alternative approach

Use the answer choices. Choice (C) 51100\frac{51}{100} and (E) 1325\frac{13}{25} are both bigger than 5099\frac{50}{99} (for example 51×99=5049>50×100=500051 \times 99 = 5049 > 50 \times 100 = 5000), so they would include the slope 5099\frac{50}{99}, which hits (99,52)(99, 52). Choice (D) 52101\frac{52}{101} is bigger still. Choice (A) 51101\frac{51}{101} is smaller than 5099\frac{50}{99}, since 51×99=5049<50×101=505051 \times 99 = 5049 < 50 \times 101 = 5050; it is safe but not the largest. That leaves (B).

Concise solution

A shorter write-up, the way contest solutions are usually written. It may take a different route from the solution above.

At an integer xx, the point (x,mx+2)(x, mx + 2) is a lattice point exactly when mxmx is an integer, i.e. when m=kxm = \frac{k}{x} for some integer kk. So the line avoids lattice points for 0<x≤1000 < x \le 100 precisely when mm is not a fraction with denominator (before or after reduction) at most 100100.

Therefore the interval (12,a)\left(\frac12, a\right) is safe as long as it contains no fraction kx\frac{k}{x} with 1≤x≤1001 \le x \le 100, and the largest possible aa is the smallest such fraction exceeding 12\frac12.

For kx>12\frac{k}{x} > \frac12 we need 2k−x≥12k - x \ge 1, so
kx−12=2k−x2x≥12x.\frac{k}{x} - \frac12 = \frac{2k - x}{2x} \ge \frac{1}{2x}.
To make this gap as small as possible we want 2k−x=12k - x = 1 (so xx must be odd) and xx as large as possible: x=99x = 99, k=50k = 50. The candidate is 5099\frac{50}{99}, with gap 1198\frac{1}{198}.

Could an even denominator do better? With x=100x = 100 the best is 2k−x=22k - x = 2, giving 51100\frac{51}{100} and gap 1100\frac{1}{100}, which is larger. Cross-check: 50⋅100=5000<51⋅99=504950 \cdot 100 = 5000 < 51 \cdot 99 = 5049, so indeed 5099<51100\frac{50}{99} < \frac{51}{100}.

Hence every mm in (12,5099)\left(\frac12, \frac{50}{99}\right) is safe, while m=5099m = \frac{50}{99} itself hits the lattice point (99,52)(99, 52). The maximum aa is 5099\frac{50}{99}.

The answer is (B) 5099\boxed{\textbf{(B)}\ \frac{50}{99}}.

Review in the AoPS books

Similar problems

Problem © Mathematical Association of America (MAA), American Mathematics Competitions. Reproduced for non-commercial educational use. Solution and commentary are original to this site.