Bank stats — CCA-F
Question inventory, coverage, and quality signals. Aggregate only — individual users' scores stay private.
Inventory
885
questions in the bank
| Verified | 872 |
| Awaiting review | 13 |
| Rejected (excluded from play) | 0 |
| Difficulty mix | 16 easy · 373 medium · 496 hard |
Usage
| Served at least once | 94 of 885 (11%) |
| Never seen by anyone | 791 |
| Study sessions | 37 |
| Answers recorded | 138 |
| Users | 10 |
| Open question flags | 0 |
By domain (bank share vs exam weight)
| Domain | Exam weight | Bank share | Verified | Unverified | Served | Miss rate |
|---|---|---|---|---|---|---|
| Agentic Architecture | 27% | 26.2% (232) | 230 | 2 | 23/232 | 34% of 32 |
| Claude Code Configuration | 20% | 20.5% (181) | 177 | 4 | 22/181 | 53% of 36 |
| Prompt Engineering | 20% | 20.7% (183) | 179 | 4 | 20/183 | 48% of 27 |
| Tool Design & MCP | 18% | 17.9% (158) | 157 | 1 | 16/158 | 50% of 24 |
| Context Management | 15% | 14.8% (131) | 129 | 2 | 13/131 | 37% of 19 |
⚠ = the bank is thinner than the exam weight suggests — generate or import more items there so mocks sample it realistically.
By author (who wrote the questions, and how they play)
| Author / model | Source | Questions | Answers seen | Miss rate |
|---|---|---|---|---|
chatgpt |
imported bank | 412 | 20 | 15% |
claude-fable-5 |
imported bank | 192 | 102 | 44% |
gemini |
imported bank | 96 | 1 | 100% |
olivier-alter-practice-exam |
imported bank | 77 | 1 | 100% |
expert-exam-author |
imported bank | 53 | 0 | — |
claude-platform-mechanics-docs |
imported bank | 24 | 0 | — |
mc-marcocheng-gist-cherrypick |
imported bank | 17 | 1 | 100% |
Qwen3.6-35B-A3B-GGUF |
generated in-app | 14 | 13 | 85% |
A source whose miss rate runs far above the rest may be writing harder — or lower-quality — questions. Cross-check with flags before judging.
Most-missed questions (min 2 answers)
| # | Question | Domain | Missed |
|---|---|---|---|
| 98 | Every task in your tracker has an ID, and company convention requires the current ticket's… | Claude Code Configuration | 5/6 (83%) |
| 47 | A high-volume LLM feature has p95 latency and cost problems. Profiling shows: 30k input to… | Context Management | 4/5 (80%) |
| 74 | You're deploying a community-built MCP server that syncs with your project-management SaaS… | Tool Design & MCP | 4/5 (80%) |
| 77 | An agent workflow frequently needs current exchange rates for a currency pair and then con… | Tool Design & MCP | 4/5 (80%) |
| 26 | The team wants every Claude Code edit to Python files to be immediately followed by runnin… | Claude Code Configuration | 3/4 (75%) |
| 54 | A team iterates on their triage prompt weekly: someone tweaks wording, tries two or three … | Prompt Engineering | 3/4 (75%) |
| 60 | A prompt asks the model to 'think step by step, then answer' for a compliance-decision tas… | Prompt Engineering | 3/4 (75%) |
| 109 | You create a 'security-auditor' subagent. Its job: scan dependency manifests and flag risk… | Claude Code Configuration | 3/4 (75%) |
| 1 | Your team is designing an autonomous software maintenance agent capable of diagnosing prod… | Agentic Architecture | 2/3 (67%) |
| 9 | Your debugging agent occasionally spirals: it re-reads the same files and retries similar … | Agentic Architecture | 2/3 (67%) |