After a session

Two screens, and they answer different questions. The analysis says how it went; the review is the questions themselves.

Address	Is	Answers
/sessions/:attemptId	The session analysis	How did that go, and what should I do now?
/results/:attemptId	The review	What did I put, and why was it wrong?

Where you land

Handing in an exam takes you to the analysis. Unless the clock ran out — a block the clock ended stays in the player, which has just switched to review with the answers on screen. Taking the paper away at the moment it finally became readable would be the wrong moment; Exit goes to the analysis.

A study session never has a hand-in moment, so it simply becomes its own review once every question has an answer.

The analysis

The four figures

Correct percent, answered of total, seconds per question, total time. They read — until the session is finished, and the page says why rather than showing a score of nought:

• an exam still running has not been marked
• a session with questions left says how many, and that sitting it through fills the page in

The ring

Correct, incorrect, unanswered. An unmarked-but-answered session shows "answered" instead of a right/wrong split, because there is not one yet.

Study recommendations

The same question asked three ways — Articles, Disciplines, Systems — using the same ranking as the standalone recommendations. What is behind that ranking is in Progress and analysis.

Performance analytics

One row per question: the stem, the status, the difficulty, your time, and peers correct with the sample size it is drawn from. Sortable in order, slowest first or hardest first, and paged.

Clicking a row lands on that question, not at the top of the session — in the review if the session is finished, in the player at that question if it is not.

A question's status is one of correct, incorrect, skipped, withdrawn, unanswered.

Repeat session

Not "start the whole thing again". It asks two things — which outcomes (not yet answered, answered incorrectly, answered correctly) and how many — and builds a session from exactly those, shuffled.

The default is incorrect and unanswered, which is usually what repeating means. It does not wait for the session to be finished: the questions worth sitting again are worth sitting again now.

The new session is always built in study mode and marked is_repetition, which means it is excluded from every performance figure — the dashboard, the history and the analytics. Re-answering questions you have already seen is practice, not a new measurement. Its own analysis page still works.

There is no delete

A session is a record of work done, and removing one edits the history every figure is computed from. Starting again is offered whole, under Settings → Your data.

The review

/results/:attemptId wears the player: the rail on the left with every question and how it went, the question in the middle, the lab column beside it. Reviewing used to be a page of its own with a different layout, so moving from sitting a block to reading it back meant learning a second set of controls for the same act.

• No score is shown here. A running mark at the top turns reading your mistakes into a report card; the score is on the analysis page.
• Lab values are available. So are notes — an exam's notes come back in review, because that is when there is something worth writing down.
• ?q=3 opens on question 3, which is how the analytics table links in.
• The tutor is always allowed here. The answer is already out.

When there is nothing to review

Two different sentences, for two different situations:

• Still open — answers are held as progress until the session is handed in, so there is nothing to read back yet. The button offers Resume.
• Handed in with no answers — there is genuinely nothing. The button offers Sit it again.

How everybody else did

Once the attempt is finished (or at any time in a study attempt), each option carries the share of recorded answers that chose it, with the sample size underneath. Show stats / Hide stats toggles it, and the choice is remembered on that device.

The distribution counts answers from completed, non-expired attempts on questions you are allowed to see. Skips and options the question no longer has are excluded. A percentage and nothing else — on a young bank, "0/1" beside an option reads as a verdict on the option rather than as a small sample.

Across every session

/sessions is the list and the picture across all of them: performance over time, the completion tiles, readiness against peers and the focus-area table. Every figure there, and exactly what it counts, is in Progress and analysis.

Throughout: only completed, non-expired, non-repetition attempts count.