Understand the failure.
Verify the repair.
Small, reproducible software failures. The broken implementation, the fix that didn’t work, and the one that passed—preserved together.
Explore the cases ↓How results are verified ↗WHAT THE ARCHIVE CONTAINS
100840 executable cases. 20168 are open.
Every case records the implementation that fails, the fix that did not work, and the repair that passed its checks—with recorded outputs and source hashes. This release adds 100840 cases across 20168 failure mechanisms and 254 domains.
The open tier gives you the failure and the unsuccessful fix for one case in every mechanism. The remaining 80672 cases, 5 variants per mechanism, are member-only: the verified repair, its recorded checks, and the full fixture suite are held in the member archive. Read the methodology ↗
A RECORD OF WHAT WENT WRONG
Browse the archive / 100840
Split exam fills smallest rooms first · case 01
Split exams fragment across many small rooms.
Split exam fills smallest rooms first · case 02
Split exams fragment across many small rooms.
Split exam fills smallest rooms first · case 03
Split exams fragment across many small rooms.
Split exam fills smallest rooms first · case 04
Split exams fragment across many small rooms.
Split exam fills smallest rooms first · case 05
Split exams fragment across many small rooms.
Rooms used by a split exam keep no duration record · case 01
A later exam of a different duration is seated in a room already running part of a split exam.
Rooms used by a split exam keep no duration record · case 02
A later exam of a different duration is seated in a room already running part of a split exam.
Rooms used by a split exam keep no duration record · case 03
A later exam of a different duration is seated in a room already running part of a split exam.
Rooms used by a split exam keep no duration record · case 04
A later exam of a different duration is seated in a room already running part of a split exam.
Rooms used by a split exam keep no duration record · case 05
A later exam of a different duration is seated in a room already running part of a split exam.
Exams placed smallest first · case 01
Large exams are left for last and fail to find rooms.
Exams placed smallest first · case 02
Large exams are left for last and fail to find rooms.
Exams placed smallest first · case 03
Large exams are left for last and fail to find rooms.
Exams placed smallest first · case 04
Large exams are left for last and fail to find rooms.
Exams placed smallest first · case 05
Large exams are left for last and fail to find rooms.
Proximity weight grows with distance · case 01
Exams far apart are penalised more than adjacent ones.
Proximity weight grows with distance · case 02
Exams far apart are penalised more than adjacent ones.
Proximity weight grows with distance · case 03
Exams far apart are penalised more than adjacent ones.
Proximity weight grows with distance · case 04
Exams far apart are penalised more than adjacent ones.
Proximity weight grows with distance · case 05
Exams far apart are penalised more than adjacent ones.
Same-period clashes included in proximity cost · case 01
Clashing pairs add 32 to the soft cost.
Same-period clashes included in proximity cost · case 02
Clashing pairs add 32 to the soft cost.
Same-period clashes included in proximity cost · case 03
Clashing pairs add 32 to the soft cost.
Same-period clashes included in proximity cost · case 04
Clashing pairs add 32 to the soft cost.
Same-period clashes included in proximity cost · case 05
Clashing pairs add 32 to the soft cost.
Repeated enrolment rows multiply proximity pairs · case 01
A duplicated enrolment doubles the cost of every pair involving that exam.
Repeated enrolment rows multiply proximity pairs · case 02
A duplicated enrolment doubles the cost of every pair involving that exam.
Repeated enrolment rows multiply proximity pairs · case 03
A duplicated enrolment doubles the cost of every pair involving that exam.
Repeated enrolment rows multiply proximity pairs · case 04
A duplicated enrolment doubles the cost of every pair involving that exam.
Repeated enrolment rows multiply proximity pairs · case 05
A duplicated enrolment doubles the cost of every pair involving that exam.
Proximity cost normalised by enrolment rows · case 01
The reported per-student cost shrinks as students take more exams.
Proximity cost normalised by enrolment rows · case 02
The reported per-student cost shrinks as students take more exams.
Proximity cost normalised by enrolment rows · case 03
The reported per-student cost shrinks as students take more exams.
Proximity cost normalised by enrolment rows · case 04
The reported per-student cost shrinks as students take more exams.
Proximity cost normalised by enrolment rows · case 05
The reported per-student cost shrinks as students take more exams.
Extra-time minutes floored before five-minute rounding · case 01
Some extended sittings end five minutes early.
Extra-time minutes floored before five-minute rounding · case 02
Some extended sittings end five minutes early.
Extra-time minutes floored before five-minute rounding · case 03
Some extended sittings end five minutes early.
Extra-time minutes floored before five-minute rounding · case 04
Some extended sittings end five minutes early.
Extra-time minutes floored before five-minute rounding · case 05
Some extended sittings end five minutes early.
Extended duration rounded down to the five-minute grid · case 01
Candidates lose up to four minutes of their allowance.
Extended duration rounded down to the five-minute grid · case 02
Candidates lose up to four minutes of their allowance.
Extended duration rounded down to the five-minute grid · case 03
Candidates lose up to four minutes of their allowance.
Extended duration rounded down to the five-minute grid · case 04
Candidates lose up to four minutes of their allowance.
Extended duration rounded down to the five-minute grid · case 05
Candidates lose up to four minutes of their allowance.
Exact rest gap flagged as a conflict · case 01
Sittings separated by exactly the required rest are reported as conflicts.
Exact rest gap flagged as a conflict · case 02
Sittings separated by exactly the required rest are reported as conflicts.
Exact rest gap flagged as a conflict · case 03
Sittings separated by exactly the required rest are reported as conflicts.
Exact rest gap flagged as a conflict · case 04
Sittings separated by exactly the required rest are reported as conflicts.
Exact rest gap flagged as a conflict · case 05
Sittings separated by exactly the required rest are reported as conflicts.
Extra-time percentage added as flat minutes · case 01
A 25% allowance on a three-hour exam adds only 25 minutes.
Extra-time percentage added as flat minutes · case 02
A 25% allowance on a three-hour exam adds only 25 minutes.
Extra-time percentage added as flat minutes · case 03
A 25% allowance on a three-hour exam adds only 25 minutes.
Extra-time percentage added as flat minutes · case 04
A 25% allowance on a three-hour exam adds only 25 minutes.
Extra-time percentage added as flat minutes · case 05
A 25% allowance on a three-hour exam adds only 25 minutes.
Extended sittings compared in input order · case 01
A student's exams listed out of order produce spurious or missed conflicts.
Extended sittings compared in input order · case 02
A student's exams listed out of order produce spurious or missed conflicts.
Extended sittings compared in input order · case 03
A student's exams listed out of order produce spurious or missed conflicts.
Extended sittings compared in input order · case 04
A student's exams listed out of order produce spurious or missed conflicts.
Extended sittings compared in input order · case 05
A student's exams listed out of order produce spurious or missed conflicts.
Invigilator count floors the candidate ratio · case 01
Rooms with fewer candidates than the ratio get no invigilator.
Invigilator count floors the candidate ratio · case 02
Rooms with fewer candidates than the ratio get no invigilator.
Invigilator count floors the candidate ratio · case 03
Rooms with fewer candidates than the ratio get no invigilator.
Invigilator count floors the candidate ratio · case 04
Rooms with fewer candidates than the ratio get no invigilator.
Invigilator count floors the candidate ratio · case 05
Rooms with fewer candidates than the ratio get no invigilator.
Empty exam rooms staffed with an invigilator · case 01
Rooms with no candidates consume invigilators.
Empty exam rooms staffed with an invigilator · case 02
Rooms with no candidates consume invigilators.
Empty exam rooms staffed with an invigilator · case 03
Rooms with no candidates consume invigilators.
Empty exam rooms staffed with an invigilator · case 04
Rooms with no candidates consume invigilators.
Empty exam rooms staffed with an invigilator · case 05
Rooms with no candidates consume invigilators.
Chief invigilator assigned to an unused venue · case 01
Venues with only empty rooms still receive a chief invigilator.
Chief invigilator assigned to an unused venue · case 02
Venues with only empty rooms still receive a chief invigilator.
Chief invigilator assigned to an unused venue · case 03
Venues with only empty rooms still receive a chief invigilator.
Chief invigilator assigned to an unused venue · case 04
Venues with only empty rooms still receive a chief invigilator.
Chief invigilator assigned to an unused venue · case 05
Venues with only empty rooms still receive a chief invigilator.
Grand total omits chief invigilators · case 01
The staffing request is short by one person per venue.
Grand total omits chief invigilators · case 02
The staffing request is short by one person per venue.
Grand total omits chief invigilators · case 03
The staffing request is short by one person per venue.
Grand total omits chief invigilators · case 04
The staffing request is short by one person per venue.
Grand total omits chief invigilators · case 05
The staffing request is short by one person per venue.
Precedence constraint satisfied by sharing a period · case 01
An exam required to follow another is accepted in the same period.
Precedence constraint satisfied by sharing a period · case 02
An exam required to follow another is accepted in the same period.
Precedence constraint satisfied by sharing a period · case 03
An exam required to follow another is accepted in the same period.
Precedence constraint satisfied by sharing a period · case 04
An exam required to follow another is accepted in the same period.
Precedence constraint satisfied by sharing a period · case 05
An exam required to follow another is accepted in the same period.
Exam exactly as long as its period flagged · case 01
Exams that exactly fill their period are rejected.
Exam exactly as long as its period flagged · case 02
Exams that exactly fill their period are rejected.
Exam exactly as long as its period flagged · case 03
Exams that exactly fill their period are rejected.
Exam exactly as long as its period flagged · case 04
Exams that exactly fill their period are rejected.
Exam exactly as long as its period flagged · case 05
Exams that exactly fill their period are rejected.
Period durations looked up with one-based indices · case 01
Each exam is checked against the previous period's length.
Period durations looked up with one-based indices · case 02
Each exam is checked against the previous period's length.
Period durations looked up with one-based indices · case 03
Each exam is checked against the previous period's length.
Period durations looked up with one-based indices · case 04
Each exam is checked against the previous period's length.
Period durations looked up with one-based indices · case 05
Each exam is checked against the previous period's length.
Exclusive exam conflicts with itself · case 01
Every exclusive-room constraint is reported even when the exam is alone.
Exclusive exam conflicts with itself · case 02
Every exclusive-room constraint is reported even when the exam is alone.
Exclusive exam conflicts with itself · case 03
Every exclusive-room constraint is reported even when the exam is alone.
Exclusive exam conflicts with itself · case 04
Every exclusive-room constraint is reported even when the exam is alone.
Exclusive exam conflicts with itself · case 05
Every exclusive-room constraint is reported even when the exam is alone.
INSPECTABLE BY DESIGN
Every result has a runnable source.
Runnable implementations with recorded outputs, source hashes, and explicit contracts. Related variants share a failure mechanism and belong together in evaluation splits.
Read the methodology ↗