Every assessment, and who cleared it.
Ranked on what the authors missed.
We rank experts on real work rather than credentials. A place on a leaderboard is earned by finding what the authors of that work missed, and by two people who sat the same exam confirming it blind.
See the leaderboards- 01
Specific
What breaks, and how to check it.
- 02
New to the work
Not already known about the work.
- 03
Confirmed twice
Two peers confirm it, blind.
Built so the result cannot be gamed.
Reviewers work blind
The author is never sent to the reviewer.
Nothing to memorise
Every problem a script could catch is published in advance.
One attempt each
An assessment cannot be practised into or resat.
Small cohorts go unranked
Under five scored experts, a board publishes with no order.
Hiring from a board.
The people on these boards are on our panel, and their result is on the record. Tell us what you need read, judged or built, and we will put the work in front of them.
Questions, answered.
What an assessment is, how a finding earns a result, and what a board does and does not put in public.
That every person on it found something real about a live piece of work which its own authors had missed, and that two other experts confirmed the finding without knowing whose it was. It is a record of work done, not a claim anybody made about themselves.
So recall is worth nothing. The list is drawn on a rule rather than by taste: if a script could catch it, it is on the list. Naming one earns no credit either way, because we already ran the script. What is left is judgement about the work, and that is the thing being measured.
Precision. Of the findings that were judged, how many came back confirmed. Ties go to whoever had more confirmed, then to whoever submitted first. Every figure a row carries is published beside it, so the order can be checked rather than taken on trust.
Not by volume, because the ranking is precision and submitting more does not move you up. Not by memorising, because everything a script could catch is published in advance. Not by resitting, because there is one attempt per person per assessment.
Two other participants, with the author removed from the record they receive rather than merely hidden on the page. Findings are never published, and there is no public archive of them.
An assessment needs two things: a real artefact, and an honest list of the problems its authors already know about. If you have both, get in touch and we will build the exam around it and run the cohort.