Skip to content

ICPC Modeling Under Pressure 01

Who This Is For

Use this drill if:

  • your team often says “we knew the algorithm later, but we read the problem badly”
  • handoffs are noisy because the model never became compact
  • the opening 20 minutes feel chaotic even when the topic knowledge is there

Entry Gate

Run this only if standard contest-core families already feel like review.

At minimum, most of these should feel recognizable:

Drill Goal

  • practice turning raw statements into one-machine-ready models fast enough that triage and handoff stay clean

Starter Block

Use this repo-native 5-problem block:

These fit Wave A because they are all anchor notes with sharp recognition cues and they force five different first models without collapsing into one giant investment problem.

Round Setup

  • time budget: 60-75 minutes
  • inputs:
  • the 5 statements above
  • one blank Opening Triage Board
  • one blank Model Brief Card for the top 3 candidates
  • output artifacts:
  • one filled triage board
  • one clean model summary for the top 3 candidates
  • one explicit first-solve choice and one backup solve

Good Repo Anchors

Running Rules

0-12 min: Cold Read

  • read all 5 problems
  • tag each as banker, medium, or investment
  • do not touch real contest code yet

12-25 min: Model Brief For The Top 3

For the top 3 candidates only, write one compact model brief:

family / primitive complexity main pitfall first test machine priority
high / medium / low

If this card cannot be filled in about 90 seconds, the problem is not machine-ready yet.

25-40 min: Pressure-Test The Choice

Ask:

  • is this still the safest first solve?
  • did one “investment” problem actually collapse into a medium or banker once modeled clearly?
  • is the machine priority based on solve safety or just excitement?

40-55 min: Commit One First Solve And One Backup

  • choose one first solve
  • choose one backup solve
  • state the first concrete coding step for both

55-70 min: Debrief

Log exactly:

  1. one problem whose first model was wrong
  2. one problem whose model was clean enough for a handoff
  3. one machine-priority decision you would keep next time

Triage Board

Use the board from ICPC Playbook.

Model Brief Card

Use this compact card for the top 3 candidates only:

family / primitive complexity main pitfall first test exact next coding step

Allowed Refreshers

Success Criteria

  • the top candidates have compact model briefs instead of vague verbal summaries
  • the machine priority is explicit
  • at least one problem changed category because the model became clearer

Debrief

Write down:

  1. first wrong triage tag
  2. first wrong primitive
  3. the problem that should have received machine priority earlier
  4. whether the failure was mainly mis-triage, wrong primitive, implementation stall, bad bailout, or bad handoff

Failure Routing

Failure signal Go next
no clean opening protocol ICPC Playbook
topic recognition is weak, not modeling Ladders
implementation trust is the real bottleneck Build Kit
switching is weak across many slots ICPC Breadth And Implementation 01

After This Drill

Research snapshot refreshed against official ICPC sources on 2026-04-24.