AI is pretty good at AMC prep strategy. On paper.

Ask it how students should prepare, and it gives you a sensible list of practice methods. Ask whether they should learn advanced math, and it tells you to value depth over breadth. Ask about time management, and it recommends a pass-based system with rules for moving on. Ask whether students should guess, and it explains educated guessing.

That all sounds useful.

But on the AMC, the hard part is rarely knowing the advice. It’s knowing when, how, and whether that advice applies to the student sitting in front of you.

That’s where AI keeps falling short. It gives the category of answer. It misses the coaching judgment.

Note: We had a parent ask Gemini these questions on May 5, 2026. The answers have likely changed since then—hopefully for the better!

Question 1: What’s the best way to prepare for the AMC?

AI Answer:

The best way to prepare is to engage in deliberate, problem-centered learning that forces the brain to construct solutions from first principles.

• Work backward: Use past exams to identify patterns in the types of problems they struggle with most.
• Focus on process: Prioritize understanding the logic of a solution rather than just getting the right answer.
• Limit resources: Avoid calculators entirely; develop mental math speed and estimation skills.
• Active discussion: Explaining a solution to another person is the best way to solidify one’s own understanding.

Wes’s Response:

This answer is reasonable, but it’s pretending there’s one best way to prepare. There isn’t.

The best way to prepare for the AMC is the way that works for the student in front of you. For one student, that might mean grinding past exams and building pattern recognition. For another, it might mean slowing down and learning how to really unpack a problem instead of grabbing the first tool that comes to mind. For another, it might mean practicing discussion, because they don’t actually know what they know until they have to explain it out loud.

So yes, this list has good ideas. Past exams, solution review, mental math, explaining your reasoning. Fine. None of that is crazy.

But it’s not a plan. It’s a buffet.

The work is figuring out which of those strategies actually changes the student’s performance. That’s the part AI skips, because AI can hand you a menu, but it can’t watch a student solve problems and say, “This is the thing getting in your way.” That’s what coaching is for.

Question 2: Should students learn more advanced math?

AI Answer:

Learning advanced math is beneficial only if it deepens the student’s understanding of core contest math principles rather than replacing them.

• Depth over breadth: Understanding how to manipulate complex algebra is useful, but the contest focuses on how those concepts appear in puzzles.
• Relevant advanced topics: Topics like modular arithmetic, basic number theory, and counting/probability are more “AMC-useful” than calculus.
• Strategic learning: Only move to advanced topics if the foundational contest math is already mastered.
• Foundation check: If a student cannot solve a geometry problem using basic theorems, learning advanced coordinate geometry will not help.

Wes’s Response:

No. The answer is no.

Students should not prepare for the AMC by learning more advanced math. I cannot tell you how many students know some fancy theorem they didn’t need in the first place and then can’t recognize when basic geometry would solve the problem cleanly.

That is not preparation. That is collecting tools because tools feel like progress.

Students need to get better at using the math they already “know.” And I mean actually know, not “have seen once,” not “can recite the theorem,” not “can use it when the worksheet title tells them exactly what to do.”

So yes, “depth over breadth” is right. But most people misunderstand what that means. It doesn’t mean learn slightly fancier math and call it depth. It means take the concepts already on the AMC and practice recognizing them in unfamiliar contexts, under pressure, when the problem is deliberately trying not to announce itself.

That is where students get better.

Question 3: Is it better to practice problems or study theory?

AI Answer:

It is almost always better to practice problems, provided those problems are used as a gateway to discovering and understanding the underlying theory.

• The “Why” factor: Encountering a problem, failing to solve it, and then learning the theory to solve it creates a much stronger mental bond.
• Active application: Theoretical study is passive; problem-solving is active and mirrors the testing environment.
• Theory as a tool: Theory should be learned “just in time” as a solution to a problem, not “just in case” by reading textbooks cover to cover.
• Contextual learning: Knowing a theorem in isolation is rarely enough; students need to see how it “hides” inside a contest question.

Wes’s Response:

This answer is correct, but vacuous.

Question 4: How should students manage time during the test?

AI Answer:

Time management is the difference between a good score and a great one; students must ruthlessly prioritize based on difficulty.

• The 3-Pass Method: * Pass 1: Solve all easy problems immediately.
  • Pass 2: Address the medium-difficulty problems.
  • Pass 3: Spend remaining time on the most difficult items.
• Identify traps: Recognize problems that look simple but have “hidden” complexity, and be willing to leave them for last.
• Stop/Drop/Move: If a student has spent 3–4 minutes on a problem with no progress, they must move on immediately.
• Efficiency: Use estimation and process of elimination to narrow down multiple-choice options when time is running out.

Wes’s Response:

For once, the AI’s answer points at something important.

“Identify traps” sounds nice, but in practice, no, students usually can’t just do that on command. That’s like saying you can learn Japanese by memorizing the characters. Technically related, but it hides so much complexity that it’s a useless bullet point.

The useful part is “stop, drop, move.” It’s not calibrated quite right, but it points at the real issue: students take problems personally.

If they can’t get a problem in the first 60 seconds, many students don’t make a decision. They just keep fighting it. Then suddenly they’re still there 12 minutes later, stuck in the mud.

The skill isn’t “move on after 3 or 4 minutes.” The skill is learning to recognize the feeling of being stuck, then making a conscious decision: stay in the mud for another 30 seconds, or move on.

Question 5: Should students guess?

AI Answer:

Whether to guess depends on the specific scoring rules of the AMC version, but generally, educated guessing is a high-value strategy.

• Understand scoring: If there is no penalty for an incorrect answer, a blind guess is better than leaving a blank.
• Elimination: Never guess without first eliminating at least two clearly incorrect options, which significantly improves the probability of success.
• Strategic guessing: Use the nature of the math to eliminate extremes (e.g., if the answer must be an integer, eliminate non-integers).
• Probability: On later problems where you have zero time, guessing is statistically better than leaving it blank, provided the test structure doesn’t penalize wrong answers.

Wes’s Response:

This answer makes guessing sound much more useful than it is.

The scoring rubric makes this completely clear to anybody who knows math, which one hopes includes the people taking the test. The only time students should guess on the AMC is when they can eliminate at least two answer choices with at least 90% certainty.

In other words: almost never.

There are exceptions. Parity analysis, for example, can often eliminate two answer choices. But it’s a dangerous tool because it only works if the student really knows what they’re doing.

It is trainable, though. And if a student knows what they’re doing, it can be pretty powerful.

---

Five questions. Four reasonable answers (and one very much not). Four places where any useful insight lives underneath the advice.

That’s the recurring problem with AI’s AMC guidance. It can tell students to practice problems, study deeply, manage time, and guess strategically. Lovely. Put it on a motivational poster and let it haunt a hallway.

But the actual work is more specific: figuring out how a student learns, what they only think they know, when they’re stuck, and whether a “strategy” is actually helping or just giving them something impressive-sounding to do.

The AMC rewards judgment under pressure. Preparing for it requires the same thing.

If your student is preparing for the AMC and you’re not sure whether their current approach is actually working, we’d love to talk.

Schedule a free consultation.