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
Central death rate to mortality rate conversion: A rate of exactly one is rejected · case 01
A group where everyone dies returns an error.
Central death rate to mortality rate conversion: A rate of exactly one is rejected · case 02
A group where everyone dies returns an error.
Central death rate to mortality rate conversion: A rate of exactly one is rejected · case 03
A group where everyone dies returns an error.
Central death rate to mortality rate conversion: A rate of exactly one is rejected · case 04
A group where everyone dies returns an error.
Central death rate to mortality rate conversion: A rate of exactly one is rejected · case 05
A group where everyone dies returns an error.
Central death rate to mortality rate conversion: Negative deaths are accepted · case 01
Corrupt negative death counts produce negative rates.
Central death rate to mortality rate conversion: Negative deaths are accepted · case 02
Corrupt negative death counts produce negative rates.
Central death rate to mortality rate conversion: Negative deaths are accepted · case 03
Corrupt negative death counts produce negative rates.
Central death rate to mortality rate conversion: Negative deaths are accepted · case 04
Corrupt negative death counts produce negative rates.
Central death rate to mortality rate conversion: Negative deaths are accepted · case 05
Corrupt negative death counts produce negative rates.
Central death rate to mortality rate conversion: Actuarial exposure adds only half the deaths · case 01
Actuarial-method rates equal the UDD rates.
Central death rate to mortality rate conversion: Actuarial exposure adds only half the deaths · case 02
Actuarial-method rates equal the UDD rates.
Central death rate to mortality rate conversion: Actuarial exposure adds only half the deaths · case 03
Actuarial-method rates equal the UDD rates.
Central death rate to mortality rate conversion: Actuarial exposure adds only half the deaths · case 04
Actuarial-method rates equal the UDD rates.
Central death rate to mortality rate conversion: Actuarial exposure adds only half the deaths · case 05
Actuarial-method rates equal the UDD rates.
Exposure to risk by age band: Death on the next birthday is counted in the lower age · case 01
Age-band deaths include people who died at the next age.
Exposure to risk by age band: Death on the next birthday is counted in the lower age · case 02
Age-band deaths include people who died at the next age.
Exposure to risk by age band: Death on the next birthday is counted in the lower age · case 03
Age-band deaths include people who died at the next age.
Exposure to risk by age band: Death on the next birthday is counted in the lower age · case 04
Age-band deaths include people who died at the next age.
Exposure to risk by age band: Death on the next birthday is counted in the lower age · case 05
Age-band deaths include people who died at the next age.
Exposure to risk by age band: Actuarial exposure is truncated at the study end · case 01
Deaths near the end of the study get too little exposure.
Exposure to risk by age band: Actuarial exposure is truncated at the study end · case 02
Deaths near the end of the study get too little exposure.
Exposure to risk by age band: Actuarial exposure is truncated at the study end · case 03
Deaths near the end of the study get too little exposure.
Exposure to risk by age band: Actuarial exposure is truncated at the study end · case 04
Deaths near the end of the study get too little exposure.
Exposure to risk by age band: Actuarial exposure is truncated at the study end · case 05
Deaths near the end of the study get too little exposure.
Exposure to risk by age band: Late entrants are exposed from the study start · case 01
Exposure includes time before entry.
Exposure to risk by age band: Late entrants are exposed from the study start · case 02
Exposure includes time before entry.
Exposure to risk by age band: Late entrants are exposed from the study start · case 03
Exposure includes time before entry.
Exposure to risk by age band: Late entrants are exposed from the study start · case 04
Exposure includes time before entry.
Exposure to risk by age band: Late entrants are exposed from the study start · case 05
Exposure includes time before entry.
Exposure to risk by age band: Deaths before the age band are counted · case 01
Deaths at younger ages inflate the age-band death count.
Exposure to risk by age band: Deaths before the age band are counted · case 02
Deaths at younger ages inflate the age-band death count.
Exposure to risk by age band: Deaths before the age band are counted · case 03
Deaths at younger ages inflate the age-band death count.
Exposure to risk by age band: Deaths before the age band are counted · case 04
Deaths at younger ages inflate the age-band death count.
Exposure to risk by age band: Deaths before the age band are counted · case 05
Deaths at younger ages inflate the age-band death count.
Exposure to risk by age band: Actuarial extension is applied to survivors · case 01
Exact-method exposure is extended for deaths.
Exposure to risk by age band: Actuarial extension is applied to survivors · case 02
Exact-method exposure is extended for deaths.
Exposure to risk by age band: Actuarial extension is applied to survivors · case 03
Exact-method exposure is extended for deaths.
Exposure to risk by age band: Actuarial extension is applied to survivors · case 04
Exact-method exposure is extended for deaths.
Exposure to risk by age band: Actuarial extension is applied to survivors · case 05
Exact-method exposure is extended for deaths.
Gompertz-Makeham survival: Gompertz integral omits the log of c · case 01
Survival probabilities are far too high.
Gompertz-Makeham survival: Gompertz integral omits the log of c · case 02
Survival probabilities are far too high.
Gompertz-Makeham survival: Gompertz integral omits the log of c · case 03
Survival probabilities are far too high.
Gompertz-Makeham survival: Gompertz integral omits the log of c · case 04
Survival probabilities are far too high.
Gompertz-Makeham survival: Gompertz integral omits the log of c · case 05
Survival probabilities are far too high.
Gompertz-Makeham survival: Hazard integral starts at the wrong age · case 01
Survival from age x is computed as if from age x+t.
Gompertz-Makeham survival: Hazard integral starts at the wrong age · case 02
Survival from age x is computed as if from age x+t.
Gompertz-Makeham survival: Hazard integral starts at the wrong age · case 03
Survival from age x is computed as if from age x+t.
Gompertz-Makeham survival: Hazard integral starts at the wrong age · case 04
Survival from age x is computed as if from age x+t.
Gompertz-Makeham survival: Hazard integral starts at the wrong age · case 05
Survival from age x is computed as if from age x+t.
Gompertz-Makeham survival: Makeham constant is not multiplied by duration · case 01
Accident-hazard contribution does not grow with duration.
Gompertz-Makeham survival: Makeham constant is not multiplied by duration · case 02
Accident-hazard contribution does not grow with duration.
Gompertz-Makeham survival: Makeham constant is not multiplied by duration · case 03
Accident-hazard contribution does not grow with duration.
Gompertz-Makeham survival: Makeham constant is not multiplied by duration · case 04
Accident-hazard contribution does not grow with duration.
Gompertz-Makeham survival: Makeham constant is not multiplied by duration · case 05
Accident-hazard contribution does not grow with duration.
Gompertz-Makeham survival: Force of mortality is reported at the starting age · case 01
The reported hazard lags by the projection duration.
Gompertz-Makeham survival: Force of mortality is reported at the starting age · case 02
The reported hazard lags by the projection duration.
Gompertz-Makeham survival: Force of mortality is reported at the starting age · case 03
The reported hazard lags by the projection duration.
Gompertz-Makeham survival: Force of mortality is reported at the starting age · case 04
The reported hazard lags by the projection duration.
Gompertz-Makeham survival: Force of mortality is reported at the starting age · case 05
The reported hazard lags by the projection duration.
Gompertz-Makeham survival: Degenerate c=1 case drops the B hazard · case 01
Constant-hazard parameterisations understate mortality.
Gompertz-Makeham survival: Degenerate c=1 case drops the B hazard · case 02
Constant-hazard parameterisations understate mortality.
Gompertz-Makeham survival: Degenerate c=1 case drops the B hazard · case 03
Constant-hazard parameterisations understate mortality.
Gompertz-Makeham survival: Degenerate c=1 case drops the B hazard · case 04
Constant-hazard parameterisations understate mortality.
Gompertz-Makeham survival: Degenerate c=1 case drops the B hazard · case 05
Constant-hazard parameterisations understate mortality.
Abridged to single-year mortality: Group probability is divided linearly · case 01
Single-year rates do not compound back to the group rate.
Abridged to single-year mortality: Group probability is divided linearly · case 02
Single-year rates do not compound back to the group rate.
Abridged to single-year mortality: Group probability is divided linearly · case 03
Single-year rates do not compound back to the group rate.
Abridged to single-year mortality: Group probability is divided linearly · case 04
Single-year rates do not compound back to the group rate.
Abridged to single-year mortality: Group probability is divided linearly · case 05
Single-year rates do not compound back to the group rate.
Abridged to single-year mortality: Groups are assumed to be five years wide · case 01
Infant and 1-4 groups trigger a false gap.
Abridged to single-year mortality: Groups are assumed to be five years wide · case 02
Infant and 1-4 groups trigger a false gap.
Abridged to single-year mortality: Groups are assumed to be five years wide · case 03
Infant and 1-4 groups trigger a false gap.
Abridged to single-year mortality: Groups are assumed to be five years wide · case 04
Infant and 1-4 groups trigger a false gap.
Abridged to single-year mortality: Groups are assumed to be five years wide · case 05
Infant and 1-4 groups trigger a false gap.
Abridged to single-year mortality: The open group is expanded into single-year rates · case 01
Output includes certain-death rates for every open-group year.
Abridged to single-year mortality: The open group is expanded into single-year rates · case 02
Output includes certain-death rates for every open-group year.
Abridged to single-year mortality: The open group is expanded into single-year rates · case 03
Output includes certain-death rates for every open-group year.
Abridged to single-year mortality: The open group is expanded into single-year rates · case 04
Output includes certain-death rates for every open-group year.
Abridged to single-year mortality: The open group is expanded into single-year rates · case 05
Output includes certain-death rates for every open-group year.
Abridged to single-year mortality: Tables are assumed to start at age zero · case 01
Adult-only tables are rejected with a gap.
Abridged to single-year mortality: Tables are assumed to start at age zero · case 02
Adult-only tables are rejected with a gap.
Abridged to single-year mortality: Tables are assumed to start at age zero · case 03
Adult-only tables are rejected with a gap.
Abridged to single-year mortality: Tables are assumed to start at age zero · case 04
Adult-only tables are rejected with a gap.
Abridged to single-year mortality: Tables are assumed to start at age zero · case 05
Adult-only tables are rejected with a gap.
Abridged to single-year mortality: Overlapping groups are accepted · case 01
Overlapping groups duplicate ages in the output.
Abridged to single-year mortality: Overlapping groups are accepted · case 02
Overlapping groups duplicate ages in the output.
Abridged to single-year mortality: Overlapping groups are accepted · case 03
Overlapping groups duplicate ages in the output.
Abridged to single-year mortality: Overlapping groups are accepted · case 04
Overlapping groups duplicate ages in the output.
Abridged to single-year mortality: Overlapping groups are accepted · case 05
Overlapping groups duplicate ages in the output.
Abridged period life table: Conversion ignores the separation factor · case 01
Groups with early deaths get the wrong nqx.
Abridged period life table: Conversion ignores the separation factor · case 02
Groups with early deaths get the wrong nqx.
Abridged period life table: Conversion ignores the separation factor · case 03
Groups with early deaths get the wrong nqx.
Abridged period life table: Conversion ignores the separation factor · case 04
Groups with early deaths get the wrong nqx.
Abridged period life table: Conversion ignores the separation factor · case 05
Groups with early deaths get the wrong nqx.
Abridged period life table: Person-years credit deaths with the complement of nax · case 01
Life expectancy is overstated for infant groups.
Abridged period life table: Person-years credit deaths with the complement of nax · case 02
Life expectancy is overstated for infant groups.
Abridged period life table: Person-years credit deaths with the complement of nax · case 03
Life expectancy is overstated for infant groups.
Abridged period life table: Person-years credit deaths with the complement of nax · case 04
Life expectancy is overstated for infant groups.
Abridged period life table: Person-years credit deaths with the complement of nax · case 05
Life expectancy is overstated for infant groups.
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 ↗