Hey Codeforces,
One thing that confused me early on: CF problem ratings aren't assigned by setters — they're crowd-sourced from contest performance. A problem rated 900 might have been solved by fewer people than an 800 in the same round, just because it appeared later in the standings. This means the raw rating is sometimes a poor proxy for actual learning difficulty, especially for beginners.
For example, some 900-rated problems require a clean observation you either see or you don't, while certain 800s are just tedious implementation with no real insight. For someone grinding Div 3/4, blindly following rating order can be discouraging or just inefficient.
I put together a curated problem sheet with this in mind: 70 problems across the 800–1000 range, ordered by what I'd call learning difficulty rather than raw CF rating. Each problem has a note explaining why it's placed where it is, and a Div 3/Div 4 badge so you know what type of contest it came from.
A preview
A few things that might be useful:
- Progress is saved via URL hash — just bookmark the link, no account needed
- There's a handle tracker: enter your CF handle and it auto-marks problems you've already AC'd
- The curation criteria page lists the 5 rules every problem must satisfy — if you think something doesn't belong, cite the rule and suggest a replacement via feedback
- 22 recommended Div 3/4 rounds are listed for when you're ready to test yourself under contest conditions
Link: cf-sheet.vercel.app
Happy to discuss the ordering choices or problem selection in the comments. If you think a specific problem is misplaced or shouldn't be there, I'm genuinely open to changes.
Полный текст и комментарии »
