# Fixed test rubrics for Weeks 1–2

A rubric is the scoring sheet. Use these weights unchanged when the week page tells you to release an unopened test. Copy the matching rubric into the test-release prompt. Don't change it after you see the task or your result.

> **What this teaches:** I can judge fresh work against a rule I fixed before I started. I don't move the bar to match the result.

## 20-minute quiz — 10 points

- Required behavior is correct for every supplied example: `0–4`
- The Java feature the quiz names is used correctly and safely: `0–3`
- Your explanation predicts one relevant edge case or failure: `0–3`

Save the task, start and stop time, category scores, and total. Put them in the exact private test file named by the week page. In the week's `progress.md` file, save only the score, the time, and a link to the proof.

## 60-minute weekly test — 100 points

- Eight trace, predict, or diagnose items at 5 points each: `0–40`
- Small write task 1, including the required behavior and a relevant edge case: `0–30`
- Small write task 2, including the required behavior and a relevant edge case: `0–30`

Pass: `75/100`.

The release prompt may pick different Java examples. It may not change the number of items, the weights, the time limit, or the pass line. Save the released task, where your answer is, start and stop time, category scores, and total. Put them in the exact private test file named by the week page.

## 45-minute Week 2 change task — 100 points

- Behavior against published examples: `0–40`
- Tests you add: `0–20`
- Code and design: `0–15`
- Data and privacy safety: `0–15`
- Explanation: `0–10`

You pass only when all three are true:

- total is at least `75/100`;
- behavior is at least `28/40`;
- there is no serious privacy bug or data-corruption bug.

## Explain it back

Explain why a fixed time limit, fixed weights, a fixed pass line, and an unopened task make the result trustworthy. Then explain why one test still cannot prove how you will do on a job.