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
Empirical cdf inclusive · case 01
Strict comparison excludes atoms at the query point.
Empirical cdf inclusive · case 02
Strict comparison excludes atoms at the query point.
Empirical cdf inclusive · case 03
Strict comparison excludes atoms at the query point.
Empirical cdf inclusive · case 04
Strict comparison excludes atoms at the query point.
Empirical cdf inclusive · case 05
Strict comparison excludes atoms at the query point.
Survival function strict · case 01
Including equal observations mismatches the strict survival convention.
Survival function strict · case 02
Including equal observations mismatches the strict survival convention.
Survival function strict · case 03
Including equal observations mismatches the strict survival convention.
Survival function strict · case 04
Including equal observations mismatches the strict survival convention.
Survival function strict · case 05
Including equal observations mismatches the strict survival convention.
Average ranks with ties · case 01
Each tied group receives its first rank rather than average rank.
Average ranks with ties · case 02
Each tied group receives its first rank rather than average rank.
Average ranks with ties · case 03
Each tied group receives its first rank rather than average rank.
Average ranks with ties · case 04
Each tied group receives its first rank rather than average rank.
Average ranks with ties · case 05
Each tied group receives its first rank rather than average rank.
Dense ranks with ties · case 01
Competition ranks leave gaps after tied groups.
Dense ranks with ties · case 02
Competition ranks leave gaps after tied groups.
Dense ranks with ties · case 03
Competition ranks leave gaps after tied groups.
Dense ranks with ties · case 04
Competition ranks leave gaps after tied groups.
Dense ranks with ties · case 05
Competition ranks leave gaps after tied groups.
Competition ranks with ties · case 01
Dense ranks fail to leave gaps after tied observations.
Competition ranks with ties · case 02
Dense ranks fail to leave gaps after tied observations.
Competition ranks with ties · case 03
Dense ranks fail to leave gaps after tied observations.
Competition ranks with ties · case 04
Dense ranks fail to leave gaps after tied observations.
Competition ranks with ties · case 05
Dense ranks fail to leave gaps after tied observations.
Gini mean difference rational · case 01
The sample range replaces average pairwise distance.
Gini mean difference rational · case 02
The sample range replaces average pairwise distance.
Gini mean difference rational · case 03
The sample range replaces average pairwise distance.
Gini mean difference rational · case 04
The sample range replaces average pairwise distance.
Gini mean difference rational · case 05
The sample range replaces average pairwise distance.
Effective sample size weights · case 01
Nominal record count ignores unequal weight concentration.
Effective sample size weights · case 02
Nominal record count ignores unequal weight concentration.
Effective sample size weights · case 03
Nominal record count ignores unequal weight concentration.
Effective sample size weights · case 04
Nominal record count ignores unequal weight concentration.
Effective sample size weights · case 05
Nominal record count ignores unequal weight concentration.
Pooled sample variance rational · case 01
Equal averaging ignores unequal degrees of freedom.
Pooled sample variance rational · case 02
Equal averaging ignores unequal degrees of freedom.
Pooled sample variance rational · case 03
Equal averaging ignores unequal degrees of freedom.
Pooled sample variance rational · case 04
Equal averaging ignores unequal degrees of freedom.
Pooled sample variance rational · case 05
Equal averaging ignores unequal degrees of freedom.
Frequency distribution entropy bits · case 01
Natural logarithms return nats under a bits contract.
Frequency distribution entropy bits · case 02
Natural logarithms return nats under a bits contract.
Frequency distribution entropy bits · case 03
Natural logarithms return nats under a bits contract.
Frequency distribution entropy bits · case 04
Natural logarithms return nats under a bits contract.
Frequency distribution entropy bits · case 05
Natural logarithms return nats under a bits contract.
Probability union with overlap · case 01
The overlapping event mass is counted twice.
Probability union with overlap · case 02
The overlapping event mass is counted twice.
Probability union with overlap · case 03
The overlapping event mass is counted twice.
Probability union with overlap · case 04
The overlapping event mass is counted twice.
Probability union with overlap · case 05
The overlapping event mass is counted twice.
Forward first difference · case 01
Subtraction order reverses the discrete derivative.
Forward first difference · case 02
Subtraction order reverses the discrete derivative.
Forward first difference · case 03
Subtraction order reverses the discrete derivative.
Forward first difference · case 04
Subtraction order reverses the discrete derivative.
Forward first difference · case 05
Subtraction order reverses the discrete derivative.
Second central difference · case 01
The center sample is not doubled.
Second central difference · case 02
The center sample is not doubled.
Second central difference · case 03
The center sample is not doubled.
Second central difference · case 04
The center sample is not doubled.
Second central difference · case 05
The center sample is not doubled.
Prefix sums inclusive · case 01
The current sample is excluded from each prefix.
Prefix sums inclusive · case 02
The current sample is excluded from each prefix.
Prefix sums inclusive · case 03
The current sample is excluded from each prefix.
Prefix sums inclusive · case 04
The current sample is excluded from each prefix.
Prefix sums inclusive · case 05
The current sample is excluded from each prefix.
Prefix products inclusive · case 01
Addition replaces multiplicative accumulation.
Prefix products inclusive · case 02
Addition replaces multiplicative accumulation.
Prefix products inclusive · case 03
Addition replaces multiplicative accumulation.
Prefix products inclusive · case 04
Addition replaces multiplicative accumulation.
Prefix products inclusive · case 05
Addition replaces multiplicative accumulation.
Trapezoidal integral unit grid · case 01
Every sample receives a full interval weight, including both endpoints.
Trapezoidal integral unit grid · case 02
Every sample receives a full interval weight, including both endpoints.
Trapezoidal integral unit grid · case 03
Every sample receives a full interval weight, including both endpoints.
Trapezoidal integral unit grid · case 04
Every sample receives a full interval weight, including both endpoints.
Trapezoidal integral unit grid · case 05
Every sample receives a full interval weight, including both endpoints.
Trapezoidal integral nonuniform grid · case 01
Interval widths are assumed to be one.
Trapezoidal integral nonuniform grid · case 02
Interval widths are assumed to be one.
Trapezoidal integral nonuniform grid · case 03
Interval widths are assumed to be one.
Trapezoidal integral nonuniform grid · case 04
Interval widths are assumed to be one.
Trapezoidal integral nonuniform grid · case 05
Interval widths are assumed to be one.
Simpson three point integral · case 01
Trapezoidal weights replace the quadratic-exact Simpson weights.
Simpson three point integral · case 02
Trapezoidal weights replace the quadratic-exact Simpson weights.
Simpson three point integral · case 03
Trapezoidal weights replace the quadratic-exact Simpson weights.
Simpson three point integral · case 04
Trapezoidal weights replace the quadratic-exact Simpson weights.
Simpson three point integral · case 05
Trapezoidal weights replace the quadratic-exact Simpson weights.
Central first derivative samples · case 01
The two-sided distance is treated as one step.
Central first derivative samples · case 02
The two-sided distance is treated as one step.
Central first derivative samples · case 03
The two-sided distance is treated as one step.
Central first derivative samples · case 04
The two-sided distance is treated as one step.
Central first derivative samples · case 05
The two-sided distance is treated as one step.
Central second derivative samples · case 01
A second derivative is divided by one step instead of squared step.
Central second derivative samples · case 02
A second derivative is divided by one step instead of squared step.
Central second derivative samples · case 03
A second derivative is divided by one step instead of squared step.
Central second derivative samples · case 04
A second derivative is divided by one step instead of squared step.
Central second derivative samples · case 05
A second derivative is divided by one step instead of squared step.
Linear interpolation rational · case 01
The start value is scaled rather than retained as the intercept.
Linear interpolation rational · case 02
The start value is scaled rather than retained as the intercept.
Linear interpolation rational · case 03
The start value is scaled rather than retained as the intercept.
Linear interpolation rational · case 04
The start value is scaled rather than retained as the intercept.
Linear interpolation rational · case 05
The start value is scaled rather than retained as the intercept.
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 ↗