My readiness
Two very different kinds of evidence, kept visually apart on purpose. Course completion is self-reported — you ticking a box. Practice results are measured from items you actually answered, and reported as percent-correct per domain. Neither is a score prediction, and nothing on this page converts one into the 720 cut.
This page is not the system of record for anything that pays money. The
professional-development bonus pays on separately verified completion of all
4 CPN courses, through the existing bonus process with a human verifier —
see docs/partner-status.md and the course table in docs/roster.md.
Ticking a box here changes your study view and nothing else: it is not a claim, not a
submission, and not evidence of payment eligibility.
Stored in this browser only. Everything below lives in this browser's
localStorage, so a different laptop or a cleared browser starts empty, and nobody
else can see it from here. The durable per-person record is the D1 database keyed on your
Access-verified email; this page reads the local cache until that endpoint lands, at which
point the same keys carry over.
The 4 CPN courses
self-reportedNo exam, no credential, and not exam readiness. On our mapping these four cover D2 (18%) + D3 (20%) = 38% of the CCAR-F blueprint outright, leaving 62% across 3 domains. The largest single domain — D1Agentic Architecture & Orchestration, worth 27% — is taught by none of them. D1 gets touched but not covered. Coverage figures are our mapping of course content onto the official domain weights, a planning estimate rather than an Anthropic-published crosswalk. The weights themselves are the guide's.
0 of 4 marked complete
Building with the Claude API
no CCAR-F domain outright
Feeds CCDV-F and CCAR-P rather than the CCAR-F blueprint. Still worth doing — it is one of the four the bonus pays on.
Introduction to Model Context Protocol
Claude Code in Action
Verification is somebody else's job by design: mark the verifier in
docs/roster.md, and let the bonus process pay on its own evidence.
Attempt history
measuredEvery practice run this browser has stored, newest first. A run with no finish time was abandoned, which is itself worth seeing. For reference, a real sitting presents 4 scenarios drawn from the bank of 6 — 60 items in 120 minutes.
No attempts stored in this browser. Start a practice run and it will show up here. If you have run practice on another machine, that history is in that browser — this page cannot see it.
| Started | Mode | Scenarios | Items | Correct | % | State |
|---|
Per-domain trend
One row per domain. Each bar in the Runs column is one run's percent-correct on the items that run actually asked from that domain, oldest on the left; a run that asked nothing from a domain contributes no bar rather than a zero. Blueprint weight sits alongside, because a weak D1 (27%) costs about 1.8× a weak D5 (15%).
Nothing to trend yet. A trend needs at least one stored run whose items can be attributed to a domain.
Some stored responses could not be attributed to a domain. That happens when an item has left the bank since the run, or when the run was recorded without a domain on the response. Those responses are counted in the overall figures above and skipped here.
| Domain | Weight | Statements | Runs | Latest | Change |
|---|
Weakest domain is a study signal, not a verdict: with 60 items over 5 domains and 30 task statements, a single run asks only a handful per domain. Two runs pointing the same way is a signal; one is noise.
What this page will never show you
A predicted scaled score. The 720 cut on the 100–1,000 scale comes from a standard-setting study on the real item pool; a percentage on our practice bank cannot be mapped onto it, and any tool that claims otherwise is guessing. Real results report pass/fail with a scaled score plus percent-correct by domain — the percent-correct part is the only part practice can imitate.
Nor a completion certificate: the tick boxes above are your own notes. Verified completion, the bonus, and the cohort count all live outside this site.
Domains and statements Distractor families Start a practice run