GratisAI learning · 4 min read · free tool
answer-sheet error ledger

See the Pattern in Your Wrong Answers

One wrong answer is a question you missed; five wrong answers of the same type are a pattern you can fix once. This reads the questions you got wrong, classifies each miss — careless, concept gap, misread, ran out of time — and ranks which fix moves your score the most. Local only; your exam answers never leave the tab.

Scan report — what this replaces
Classification
Each miss matched to a mistake type by keyword, in your tab
local
What it replaces
A tutor's post-exam debrief of what went wrong
$0.00
Input
The questions you got wrong, one per line
no card
Output
A tally with bars and a ranked fix plan
recomputed

What you pay otherwise

RouteWhat you getCost
Tutor debriefReviews the test with you and names the pattern$40–80/hr
Re-taking the same testNew score, same unseen patterntime lost
Generic 'study more'No type, no rank, no planspread too thin
This pageYour misses grouped, tallied, and ranked$0

The workflow

  1. Paste every question you got wrong, one per line

    Include the topic or a few words of the question — 'unit rates word problem', 'quadratic with a negative sign'. The classifier keys off words like 'forgot', 'ran out of', 'misread', 'sign', 'unit'.

  2. Read the tally before you argue with the labels

    The bars show how much of your loss is one type. A single mislabel doesn't change the picture — the pattern is the bars, not any one row. You can re-word a line and re-run to sharpen it.

  3. Fix the biggest bar first

    Careless losses fix with a re-check routine; concept gaps fix with targeted re-study; time losses fix with timed practice. The plan ranks them because the biggest bar is the cheapest points to recover.

  4. Re-audit after the next test

    Paste the new misses and compare tallies. The bar that shrinks proves the fix worked; the bar that doesn't is where you look next. The audit only pays off if it's repeated.

Worked example

Sample: 8 wrong answers across a maths paper
  Concept gap       4  (50%)
  Careless          2  (25%)
  Misread           1  (12%)
  Time              1  (12%)
  Ranked fix        re-study the concept-gap topic, then a re-check routine
  Uploaded          nothing — tallied in the tab

Numbers above are sample data produced by the tool's own pre-loaded example, run in the browser — not averages or measured benchmarks.

What breaks this

Pasting only the scores, not the questions gives the classifier nothing to read — paste the question or topic, not 'Q3: 0/5'. Guessing the type when you don't remember is fine, but mark it — the tally is only as honest as the input. Fixing everything at once is the classic mistake the plan exists to prevent; one bar at a time.

Try it now

Eight wrong answers are loaded and tallied into mistake types. Paste your own misses and watch the bars re-rank.

Open the mistake audit →

Go deeper: the mistake log

The audit is free forever. The $4 one-time tier adds:

See the paid tier inside the tool →