CALE · Competency-Aligned Learning Engine

Proof that someone actually learned it.

The measurement engine behind Reaves Academy, running live in your browser. It turns a stream of answers into a defensible mastery score — asymmetric, so mastery is earned and can be lost, and time-aware, so it decays like memory does. Every constant below is read from the XML node, not hardcoded. Patent pending.

Try it right now — no signup. You are the student. Tap a few answers below (right or wrong, fast or slow) and watch a real mastery score form, cross the 80 line, then decay like memory. This is the live engine, judging your taps by the same rules it judges a classroom.
0
not started
080 · mastery line100

Answer as the learner — the engine scores each tap by whether it was right and how long it took:

Correct answer

Wrong answer

each answer updates the scoreΔ

Durability — forgetting curve

Best score reached: 0 · half-life — days

Score today: 0 ·

The contract

IN  { attempts: [ { isCorrect, responseTimeMs } ] }

OUT { smartScore 0–100, masteryBand, errorType, halfLifeDays }

Free · up to 100 subjects Embedded · $0.50 / subject / mo Platform · $150 / mo Enterprise · quote
Get access — $150/mo, Platform tier →

The engine above is the real one, not a mock

What it does

When a learner gets something wrong, CALE tells you why — not just that they did. Answering fast and wrong is a different problem from answering slowly and wrong: one is carelessness, the other is a gap they never filled. Most systems see one wrong answer and make the next question easier, which is the right move for neither. CALE names which of three things happened and what to do about it: they rushed it, they half-learned it, or they never understood it — and each one needs a different response from you.

How you use it

You do not upload your curriculum. There is nowhere to upload it to — CALE stores no content and never sees your questions. It measures answers, and your material stays exactly where it already lives.

  1. Mark each question easy or hard. One label, once. If you already have questions, they are ready as they are.
  2. Send three things when a learner answers: whether they got it right, how long they took, and which of your questions it was.
  3. Act on what comes back — the kind of mistake it was, the learner's standing, and when that idea should come round again.

If you are starting from scratch

You do not need a course to begin — you need questions with right answers. The guide in your download walks through writing them so the three kinds of mistake stay distinguishable, which is the only thing CALE asks of your content.

This is the real engine, not a mock: the code running above is the same pure library used in production, and it refuses to run if the node's constants are missing. Interested in embedding proof-of-learning? [email protected] · part of Reaves Labs.