Check How Much Text a Cloze Exercise Removes — Free
Deleting too much context can turn a language exercise into guessing. Calculate the missing-word share for each passage before deciding whether its clues still support the intended skill.
The proof surface
Language mini-drill builders generate practice. This audits deletion density in an existing cloze exercise; it neither writes answers nor assigns a proficiency level.
Why the flat version breaks
Counting blanks rather than tokens
One blank can replace several words.
Using remaining words as the base
The denominator must include the deleted text.
Equating density with difficulty
Word choice and context matter more than one ratio.
How to work the deletion stencil
Count total and deleted tokens using the same convention for punctuation and contractions. A multiword blank can hide several words, so blank count alone is not sufficient.
Keep the original passage locally so the denominator includes removed words. Do not use the word count of the remaining visible text.
Eighteen removed words out of 120 is fifteen percent. Each row gets its own percentage. The pooled summary divides all deleted words by all original words, so a short passage does not count as much as a long one.
Check whether grammar and meaning provide useful clues. The chosen threshold flags review only; a low deletion share can still remove every crucial context word.
What the deletion stencil replaces
| Question | Before | Visible working |
|---|---|---|
| Counting blanks rather than tokens | One blank can replace several words. | Count before replacing the words |
| Using remaining words as the base | The denominator must include the deleted text. | Divide deleted words by all words |
| Equating density with difficulty | Word choice and context matter more than one ratio. | Read the exercise without the key |
A local arithmetic aid replaces hand calculation, not expert review. No paid AI service is needed.
Run it on the samples, right here
FIRST-LOAD
Input is processed locally and a draft is saved automatically in this browser profile when storage is available. Reset to sample clears that draft; Pro history has its own clear button. A state link encodes your inputs in its URL: share only non-sensitive rows. Browser history, clipboard and anyone receiving the link may retain it. Optional AI use below leaves this device; it is not required.
Exercise-review arithmetic, not a validated difficulty measure or language assessment. Review the actual passage with a teacher or learner.
Data note: Everything runs in this browser tab on the deletion stencil: your passage | total word tokens | deleted word tokens stays on this device, nothing is uploaded, and the reading is rebuilt only when you press run.
Go deeper: the companion app files the same reading as a triage clipboard with three lanes
The article demo above runs without limits. The companion app keeps a local history, exports the rows as CSV, prints the deletion stencil reading, and holds your drafts on this device — one free run, then $ 4 one-time for the layer that keeps filing.
The deletion stencil reading is complete for free. The optional $4 layer adds print, row CSV and the last five local reading summaries; it does not add hidden answers. Checkout is not configured yet; the article demo remains unlimited.
Open the check how much text a cloze exercise removes companionBoundary
Exercise-review arithmetic, not a validated difficulty measure or language assessment. Review the actual passage with a teacher or learner.
What this is built on
- Method: Eighteen removed words out of 120 is fifteen percent. Each row gets its own percentage. The pooled summary divides all deleted words by all original words, so a short passage does not count as much as a long one.
- All sample records are invented examples. The app does not fetch measurements, policies, files or personal records.
- Google’s official pricing page, fetched 2026-09-27, lists free-tier access including AI Studio. Optional formatting only; limits and availability can change. Free-tier content may be used to improve products. Never paste private records there.
Before: One blank can replace several words. After: the deletion stencil shows the working beside each named row so the reader can change the assumption and inspect the consequence.
Optional AI formatting, not calculation
For this deletion stencil, use the free AI Studio interface only if available to you, with no paid API key. Manual entry completes the same workflow for free. Supply only fictional or non-sensitive notes. Review its output against the source; never paste unresolved questions into the numeric rows.
Format my non-sensitive notes for a deletion stencil. Return plain rows only: passage | total word tokens | deleted word tokens. Preserve supplied quantities exactly. Do not guess missing values; list questions separately. Do not calculate or add advice. I will check every row before pasting into the local tool.
Accepted schema: passage | total word tokens | deleted word tokens. No AI response is executed as code.