H HiMCM Prep

Six-month study plan

A week-by-week ramp from late May 2026 to the contest weekend (Nov 4–17, 2026). Assumes 3–5 hours per week during the school year, more during summer. Adjust based on Barry's real schedule.

The 3-phase logic. June–August is foundation. September–October is integration — doing increasingly contest-realistic mock papers. Late October is dress rehearsal and rest. The goal in November is to do nothing new; show up with established habits.

Phase 1 · Foundation (June – August)

Month 1 (June) · Orient and survey

Month 2 (July) · Build the modeling toolkit

Month 3 (August) · First full paper

  • Week 9. Monte Carlo, Markov chains, networkx. Implement small sensitivity-analysis demos.
  • Week 10. Read 2023-B (E-buses) in depth. Sketch a 2-page outline of how you'd attack it.
  • Week 11–12. First full mock — pick Mock M1 or M2. 4 days of work, ~5 hours/day. Write the full 15–25 page paper. Don't read the solution sketch until done.
End-of-Phase-1 checkpoint. Self-score Mock 1 against the rubric. Identify the 2 weakest sections; those become the next month's focus.

Phase 2 · Integration (September – October)

Month 4 (September) · Polish the weak spots

  • Week 13. Whichever section scored lowest in Mock 1: re-read the relevant guide, re-do the section using a past problem as practice.
  • Week 14. Read 2 more past problems carefully. Re-read the judges' commentary with fresh eyes — you'll see things you missed in June.
  • Week 15. Practice writing executive summaries: take a past problem, write a 1-page summary for it as if you'd just done the paper. Compare with what Outstanding teams wrote.
  • Week 16. Practice non-technical letters: same drill but for the 1-page memo / blog / article.

Month 5 (October) · Realistic mocks

  • Week 17. Mock 2 — pick a different style (if M1 was multi-criteria, do M3 or M4). Block off a long weekend. Treat it like the contest: no outside help, no posting online.
  • Week 18. Debrief Mock 2. Identify process problems (time management, who-did-what, getting stuck).
  • Week 19. Tighten Python toolkit. Build one reusable utility module (TOPSIS, EWM, sensitivity helpers) so you don't have to write that code under contest pressure.
  • Week 20. Test your LaTeX/Word template with a 25-page sample. Page headers, fonts, figure captions all working.

Phase 3 · Dress rehearsal & contest (late October – November 17)

Late October · Dress rehearsal

  • Week 21. Mock 3 (the dress rehearsal) — pick the hardest unused mock (M5?). 4 days, full team, contest rules. Submit a real PDF.
  • Week 22. Final debrief. Update the team's pre-flight checklist (see below).

November 1–3 · Pre-flight

  • Advisor confirms team is registered ($100 fee paid).
  • Team agrees on time blocks across the 14-day window (when each person is available).
  • Repo / shared drive / template all set up.
  • Backup tools tested (offline LaTeX install, datasets bookmarked).
  • Rest. Don't try to learn anything new. Stop adding to the toolkit.

Nov 4 (Wed) · Window opens — Day 1

  • 3:01 pm EST: advisor logs in, downloads both problems.
  • Read both problems carefully, silently, individually. ~30 min.
  • Discuss as a team. ~1 hour. Vote on A vs. B.
  • Sketch a 1-page approach for the chosen problem. Identify subtasks and assign.
  • Don't start writing the paper yet. Spend Day 1 understanding the problem.

Nov 5–13 (8 days) · Modeling phase

  • Daily 30-min standup (in person or async): what each person did yesterday, plan for today, blockers.
  • Days 2–3: assumptions, literature search, parameter values, baseline model.
  • Days 4–6: model application to scenarios, debugging.
  • Day 7: extensions and sensitivity analysis.
  • Day 8: re-read the prompt against your current draft — what requirement haven't you addressed?

Nov 14–16 · Writing & polish

  • Day 11: full first draft, all sections present (even if some are stubs).
  • Day 12: revision pass. Cut filler. Add missing citations. Make every figure earn its space.
  • Day 13: writing the Executive Summary last, after the paper is locked.

Nov 17 (Tue) · Submission

  • Morning: final proofread, anonymity check (no names in metadata!).
  • Early afternoon: export final PDF, verify pages 1–25, file size ≤ 20 MB.
  • Submit by 5:00 pm EST (4 hours before deadline) — don't be the team scrambling at 8:55.
  • Advisor receives confirmation; saves it.
  • The team takes the rest of the night off. You earned it.

Pre-flight checklist (print and tape to wall)

☐Registered before 2:00 pm EST Nov 4, $100 paid.
☐Advisor login works.
☐Team Control Number written down somewhere reliable.
☐Shared workspace (GitHub repo / Drive) set up.
☐LaTeX or Word template with 12pt font, page headers, ready.
☐Python environment pinned in requirements.txt.
☐Reusable utility code committed (TOPSIS, EWM, sensitivity).
☐Each teammate's available hours mapped onto the 14 days.
☐Communication channel agreed (team only — no external comms about the problem).
☐AI policy decision: will we use AI, and how will we log it?
☐Plan for handling disagreements about which problem to pick.

Time-allocation principle

ActivityShare of total time
Reading + understanding the problem~10%
Research / data gathering~15%
Modeling (code + math)~30%
Writing~30%
Revising + polishing~10%
Sensitivity / extensions~5%

First-time teams systematically over-allocate to modeling and under-allocate to writing. Push back on that.

Next: Resources → Or pick a mock problem