IOI Checkpoint / Score-Path Drill 01¶
Goal:
- practice banking a real safe score path before gambling on the full solution
Time budget:
90-110 minutes
Entry gate:
- you already know the core topic family well enough that the main bottleneck is proof, checkpointing, or subtask planning
Inputs:
- one official task, olympiad-style problem, or one repo note with visible easier routes
- one blank IOI Subtask-First Worksheet
Output artifacts:
- one filled worksheet
- one bank / upgrade / park decision
- one short debrief naming the first false assumption
Use This When¶
- you keep seeing the full idea too late
- you often lose safe partial points while chasing the full route
- your post-contest notes say “I almost had it” without identifying which checkpoint was actually safe
Good Repo Anchors¶
If none of those fit the family you want, take one official task and keep the same protocol.
Session Script¶
0-10 min: Read And Fill The Buckets¶
- read the statement and visible subtask or score structure
- fill every row of the IOI Subtask-First Worksheet without coding
- write the smallest valid algorithm for each visible bucket before thinking about the full solution
10-20 min: Choose The Safest Score Path¶
- select one current target bucket
- complete only the worksheet's
Current Target Proof Gate - if that proof gate still feels vague, downshift to the safest bucket you can actually prove
20-60 min: Implement Only The Checkpoint¶
- code the chosen score path only
- do not “half-build” the full solution during this phase
- keep one compiling checkpoint at all times
60-75 min: Validate At The Boundary¶
- test the smallest legal cases
- test the exact subtask boundary that justifies the chosen bucket
- if relevant, use Stress Testing Workflow or Local Judge Workflow
75-90 min: Decide Bank / Upgrade / Park¶
bankif the checkpoint is proved and stableupgradeonly if the next bucket is adjacent and the proof delta is concreteparkif the next route is still intuition-only or validation is unstable
90-100 min: Debrief¶
Log exactly:
- highest proven score path
- safe checkpoint actually banked
- first false assumption
- next repo door
Success Criteria¶
- you banked one route that you could explain and validate cleanly
- your next action is explicit:
upgrade,topic review,validation fix, orpark - the first false assumption was caught before it consumed the entire session
Failure Routing¶
| Failure signal | Go next |
|---|---|
| proof still fuzzy | Reasoning |
| validation is the real bottleneck | Stress Testing Workflow or Local Judge Workflow |
| implementation blocked a known idea | Build Kit |
| topic depth was weaker than expected | the corresponding Topic page, then the matching Ladder |
After This Drill¶
- return to IOI / OI Playbook if the track logic still fits
- reopen Problem Finder if you want the next official-style rep
- go back to Mixed Rounds only if the real bottleneck is switching rather than proof discipline
Research snapshot refreshed against official IOI sources on 2026-04-24.