Skip to main content

The result and the score

A run ends in a result word and a score. They answer two different questions.

The result​

  • PASS: no critical and no warning was found.
  • REVIEW: warnings only. Look at them before you send the file.
  • FAILED: at least one critical problem was found.
  • INCONCLUSIVE: too little of the file was checked to call it, usually because 100% matches or locked segments were skipped, so nothing found means nothing.

The result comes from the rules. A critical finding fails the file whatever the score, and a run whose AI review failed on a clean file reads PASS on rules, AI review failed in amber rather than a quiet pass.

FAILED is the file's result, not the run's: the check completed and found errors. A run that could not complete says so, names the reason and produces no findings.

The score​

A number out of 100 the tool calculates from the findings and their severities, shown as Quality 95.5/100. Critical findings cost the most; minor ones cost nothing unless the workspace gives them a weight. A rule finding the AI disputed counts at half weight until you decide. Spelling findings never count.

The score sits beside the result, not inside it: a file can read FAILED, 2 critical and Quality 95.5/100 at once, which means a good file with two things that must be fixed. Click the score to see how it was worked out: when locked, matched or another role's segments were left out, it says over how many of the file's words it was computed and where the other words are.

Useful for tracking; not a substitute for a human review.

The pass mark​

Each profile has a pass mark, under Basics. The report header shows the score against it. The workspace default is set once in Settings, Quality and a profile can set its own.

Which findings count​

In a Dropbox folder run the folder name sets the role, and Checks, Which findings count in the profile says which segments that role answers for. A segment outside that role's remit is left out: its findings are not in the report or the score, and the segment is sent to no AI review, the whole-file review included, so it costs no AI words. The one exception is a finding that shows an inconsistency with another segment: it is listed at its usual severity and tagged segment not checked, because the segment itself was not checked. It is not sent to the AI either. A file uploaded in the app has no role: it is judged as the editor's work, and a finding outside that remit is demoted to minor and kept off the score, or removed when the profile is set to leave such findings out. Evaluate everything, under Advanced, turns the remit off for both: every finding counts.

Under it, the profile's skip boxes decide which matches are not checked at all: 100% (exact matches), ICE (in-context exact matches, also called 101% or context matches) and Locked. They are separate, so a profile can skip ICE and still check 100% matches. The lines under the boxes say what each match type gets in a run.

Decisions move the score​

Ignore and Fixed on a finding take it out of the open count. The open critical count is what holds a sign off back.