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
K-index and Total Totals: cross totals moisture · case 01
Total Totals ignores low-level moisture.
K-index and Total Totals: cross totals moisture · case 02
Total Totals ignores low-level moisture.
K-index and Total Totals: cross totals moisture · case 03
Total Totals ignores low-level moisture.
K-index and Total Totals: cross totals moisture · case 04
Total Totals ignores low-level moisture.
K-index and Total Totals: cross totals moisture · case 05
Total Totals ignores low-level moisture.
K-index and Total Totals: isolated band edge · case 01
K = 25 is upgraded to widely scattered storms.
K-index and Total Totals: isolated band edge · case 02
K = 25 is upgraded to widely scattered storms.
K-index and Total Totals: isolated band edge · case 03
K = 25 is upgraded to widely scattered storms.
K-index and Total Totals: isolated band edge · case 04
K = 25 is upgraded to widely scattered storms.
K-index and Total Totals: isolated band edge · case 05
K = 25 is upgraded to widely scattered storms.
K-index and Total Totals: severe threshold · case 01
TT = 50 does not raise the severe flag.
K-index and Total Totals: severe threshold · case 02
TT = 50 does not raise the severe flag.
K-index and Total Totals: severe threshold · case 03
TT = 50 does not raise the severe flag.
K-index and Total Totals: severe threshold · case 04
TT = 50 does not raise the severe flag.
K-index and Total Totals: severe threshold · case 05
TT = 50 does not raise the severe flag.
Antecedent precipitation index: multi-day gap decay · case 01
After a week without reports the index has barely decayed.
Antecedent precipitation index: multi-day gap decay · case 02
After a week without reports the index has barely decayed.
Antecedent precipitation index: multi-day gap decay · case 03
After a week without reports the index has barely decayed.
Antecedent precipitation index: multi-day gap decay · case 04
After a week without reports the index has barely decayed.
Antecedent precipitation index: multi-day gap decay · case 05
After a week without reports the index has barely decayed.
Antecedent precipitation index: same-day accumulation · case 01
Two gauge reports for the same day decay the first report.
Antecedent precipitation index: same-day accumulation · case 02
Two gauge reports for the same day decay the first report.
Antecedent precipitation index: same-day accumulation · case 03
Two gauge reports for the same day decay the first report.
Antecedent precipitation index: same-day accumulation · case 04
Two gauge reports for the same day decay the first report.
Antecedent precipitation index: same-day accumulation · case 05
Two gauge reports for the same day decay the first report.
Antecedent precipitation index: chronological ordering · case 01
Backfilled late reports make the API grow instead of decay.
Antecedent precipitation index: chronological ordering · case 02
Backfilled late reports make the API grow instead of decay.
Antecedent precipitation index: chronological ordering · case 03
Backfilled late reports make the API grow instead of decay.
Antecedent precipitation index: chronological ordering · case 04
Backfilled late reports make the API grow instead of decay.
Antecedent precipitation index: chronological ordering · case 05
Backfilled late reports make the API grow instead of decay.
Antecedent precipitation index: sensor glitch clamp · case 01
A negative tipping-bucket glitch lowers soil moisture.
Antecedent precipitation index: sensor glitch clamp · case 02
A negative tipping-bucket glitch lowers soil moisture.
Antecedent precipitation index: sensor glitch clamp · case 03
A negative tipping-bucket glitch lowers soil moisture.
Antecedent precipitation index: sensor glitch clamp · case 04
A negative tipping-bucket glitch lowers soil moisture.
Antecedent precipitation index: sensor glitch clamp · case 05
A negative tipping-bucket glitch lowers soil moisture.
Vector-mean wind and steadiness: atan2 argument order · case 01
Westerly winds are reported as southerly.
Vector-mean wind and steadiness: atan2 argument order · case 02
Westerly winds are reported as southerly.
Vector-mean wind and steadiness: atan2 argument order · case 03
Westerly winds are reported as southerly.
Vector-mean wind and steadiness: atan2 argument order · case 04
Westerly winds are reported as southerly.
Vector-mean wind and steadiness: atan2 argument order · case 05
Westerly winds are reported as southerly.
Vector-mean wind and steadiness: compass wrap · case 01
North-westerly winds are reported with negative bearings.
Vector-mean wind and steadiness: compass wrap · case 02
North-westerly winds are reported with negative bearings.
Vector-mean wind and steadiness: compass wrap · case 03
North-westerly winds are reported with negative bearings.
Vector-mean wind and steadiness: compass wrap · case 04
North-westerly winds are reported with negative bearings.
Vector-mean wind and steadiness: compass wrap · case 05
North-westerly winds are reported with negative bearings.
Vector-mean wind and steadiness: steadiness denominator · case 01
Steadiness falls as more samples are averaged even when the wind is constant.
Vector-mean wind and steadiness: steadiness denominator · case 02
Steadiness falls as more samples are averaged even when the wind is constant.
Vector-mean wind and steadiness: steadiness denominator · case 03
Steadiness falls as more samples are averaged even when the wind is constant.
Vector-mean wind and steadiness: steadiness denominator · case 04
Steadiness falls as more samples are averaged even when the wind is constant.
Vector-mean wind and steadiness: steadiness denominator · case 05
Steadiness falls as more samples are averaged even when the wind is constant.
Vector-mean wind and steadiness: degree to radian conversion · case 01
Directions from different sensors average to meaningless values.
Vector-mean wind and steadiness: degree to radian conversion · case 02
Directions from different sensors average to meaningless values.
Vector-mean wind and steadiness: degree to radian conversion · case 03
Directions from different sensors average to meaningless values.
Vector-mean wind and steadiness: degree to radian conversion · case 04
Directions from different sensors average to meaningless values.
Vector-mean wind and steadiness: degree to radian conversion · case 05
Directions from different sensors average to meaningless values.
Aviation flight category: ceiling cover types · case 01
A scattered layer at 400 ft makes a clear day LIFR.
Aviation flight category: ceiling cover types · case 02
A scattered layer at 400 ft makes a clear day LIFR.
Aviation flight category: ceiling cover types · case 03
A scattered layer at 400 ft makes a clear day LIFR.
Aviation flight category: ceiling cover types · case 04
A scattered layer at 400 ft makes a clear day LIFR.
Aviation flight category: ceiling cover types · case 05
A scattered layer at 400 ft makes a clear day LIFR.
Aviation flight category: lowest qualifying layer · case 01
When a report lists a higher deck first, the lower ceiling is ignored.
Aviation flight category: lowest qualifying layer · case 02
When a report lists a higher deck first, the lower ceiling is ignored.
Aviation flight category: lowest qualifying layer · case 03
When a report lists a higher deck first, the lower ceiling is ignored.
Aviation flight category: lowest qualifying layer · case 04
When a report lists a higher deck first, the lower ceiling is ignored.
Aviation flight category: lowest qualifying layer · case 05
When a report lists a higher deck first, the lower ceiling is ignored.
Aviation flight category: worse-of combination · case 01
Fog with a high ceiling is reported VFR.
Aviation flight category: worse-of combination · case 02
Fog with a high ceiling is reported VFR.
Aviation flight category: worse-of combination · case 03
Fog with a high ceiling is reported VFR.
Aviation flight category: worse-of combination · case 04
Fog with a high ceiling is reported VFR.
Aviation flight category: worse-of combination · case 05
Fog with a high ceiling is reported VFR.
Aviation flight category: MVFR ceiling upper edge · case 01
A 3000 ft ceiling is reported VFR.
Aviation flight category: MVFR ceiling upper edge · case 02
A 3000 ft ceiling is reported VFR.
Aviation flight category: MVFR ceiling upper edge · case 03
A 3000 ft ceiling is reported VFR.
Aviation flight category: MVFR ceiling upper edge · case 04
A 3000 ft ceiling is reported VFR.
Aviation flight category: MVFR ceiling upper edge · case 05
A 3000 ft ceiling is reported VFR.
Aviation flight category: IFR visibility edge · case 01
Exactly 3 statute miles is reported IFR.
Aviation flight category: IFR visibility edge · case 02
Exactly 3 statute miles is reported IFR.
Aviation flight category: IFR visibility edge · case 03
Exactly 3 statute miles is reported IFR.
Aviation flight category: IFR visibility edge · case 04
Exactly 3 statute miles is reported IFR.
Aviation flight category: IFR visibility edge · case 05
Exactly 3 statute miles is reported IFR.
Pasquill-Gifford stability class: wind class boundary · case 01
A 5.0 m/s wind is assigned the lighter-wind class.
Pasquill-Gifford stability class: wind class boundary · case 02
A 5.0 m/s wind is assigned the lighter-wind class.
Pasquill-Gifford stability class: wind class boundary · case 03
A 5.0 m/s wind is assigned the lighter-wind class.
Pasquill-Gifford stability class: wind class boundary · case 04
A 5.0 m/s wind is assigned the lighter-wind class.
Pasquill-Gifford stability class: wind class boundary · case 05
A 5.0 m/s wind is assigned the lighter-wind class.
Pasquill-Gifford stability class: night cloud threshold · case 01
Four-octa nights are treated as clear and too stable.
Pasquill-Gifford stability class: night cloud threshold · case 02
Four-octa nights are treated as clear and too stable.
Pasquill-Gifford stability class: night cloud threshold · case 03
Four-octa nights are treated as clear and too stable.
Pasquill-Gifford stability class: night cloud threshold · case 04
Four-octa nights are treated as clear and too stable.
Pasquill-Gifford stability class: night cloud threshold · case 05
Four-octa nights are treated as clear and too stable.
Pasquill-Gifford stability class: overcast neutral override · case 01
Overcast days receive unstable daytime classes.
Pasquill-Gifford stability class: overcast neutral override · case 02
Overcast days receive unstable daytime classes.
Pasquill-Gifford stability class: overcast neutral override · case 03
Overcast days receive unstable daytime classes.
Pasquill-Gifford stability class: overcast neutral override · case 04
Overcast days receive unstable daytime classes.
Pasquill-Gifford stability class: overcast neutral override · case 05
Overcast days receive unstable daytime classes.
Pasquill-Gifford stability class: unknown insolation · case 01
An unrecognised insolation code silently yields a slight-insolation class.
Pasquill-Gifford stability class: unknown insolation · case 02
An unrecognised insolation code silently yields a slight-insolation class.
Pasquill-Gifford stability class: unknown insolation · case 03
An unrecognised insolation code silently yields a slight-insolation class.
Pasquill-Gifford stability class: unknown insolation · case 04
An unrecognised insolation code silently yields a slight-insolation class.
Pasquill-Gifford stability class: unknown insolation · case 05
An unrecognised insolation code silently yields a slight-insolation class.
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 ↗