Hi Codeforces! A while ago I shared LaddersCF, a free tracker for the A2OJ ladders. The response was better than I expected — thanks to everyone who tried it and signed up for updates. The big update is now live: codeforcesladders.web.app
What's new since the first post:
From 22 to 85 ladders. The classic A2OJ ones are still there (now corrected: duplicates removed, re-sorted by real problem rating — the "≥2200" ladder used to open with three 800-rated problems), plus generated ladders per rating band and per tag, reaching problems published after A2OJ froze in 2017. Per-problem difficulty tiers inside each ladder. Raw solve counts are misleading — they correlate with problem age (Spearman −0.3 to −0.5 within a rating band) because the userbase grew — so counts are normalised against same-rating, same-year peers before ranking. A new training analysis page: median solved rating over time, share of problems solved with zero wrong submissions before the first AC (per rating band), practice rhythm, and where you stall in contests versus your practice level. A coach: what to solve next (study mode vs contest-pressure mode), your weak topics measured against your own baseline, and the problems from your rated contests you never went back to finish. Faster and lighter: initial load cut from ~600 KB to ~160 KB, and it works offline as a PWA with ~3,880 statements cached. The part I care most about: nothing is vibes. The stall-point and recommendation logic comes from analysing 596 real users (100 per rating band). Example finding: 76% of users rated 1500–1699 stall at problem D. The app shows you where you sit against that.
It's still a hobby project and surely has rough edges — bug reports and feedback are very welcome, here in the comments. Hope it's useful!
