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
Quasi Newton update admits nonpositive curvature · case 01
Quasi Newton update admits nonpositive curvature.
Quasi Newton update admits nonpositive curvature · case 02
Quasi Newton update admits nonpositive curvature.
Quasi Newton update admits nonpositive curvature · case 03
Quasi Newton update admits nonpositive curvature.
Quasi Newton update admits nonpositive curvature · case 04
Quasi Newton update admits nonpositive curvature.
Quasi Newton update admits nonpositive curvature · case 05
Quasi Newton update admits nonpositive curvature.
Branch and bound prunes against an infeasible incumbent · case 01
Branch and bound prunes against an infeasible incumbent.
Branch and bound prunes against an infeasible incumbent · case 02
Branch and bound prunes against an infeasible incumbent.
Branch and bound prunes against an infeasible incumbent · case 03
Branch and bound prunes against an infeasible incumbent.
Branch and bound prunes against an infeasible incumbent · case 04
Branch and bound prunes against an infeasible incumbent.
Branch and bound prunes against an infeasible incumbent · case 05
Branch and bound prunes against an infeasible incumbent.
Simplex leaving row includes nonlimiting coefficients · case 01
Simplex leaving row includes nonlimiting coefficients.
Simplex leaving row includes nonlimiting coefficients · case 02
Simplex leaving row includes nonlimiting coefficients.
Simplex leaving row includes nonlimiting coefficients · case 03
Simplex leaving row includes nonlimiting coefficients.
Simplex leaving row includes nonlimiting coefficients · case 04
Simplex leaving row includes nonlimiting coefficients.
Simplex leaving row includes nonlimiting coefficients · case 05
Simplex leaving row includes nonlimiting coefficients.
Conjugate gradient search direction uses the residual delta · case 01
Conjugate gradient search direction uses the residual delta.
Conjugate gradient search direction uses the residual delta · case 02
Conjugate gradient search direction uses the residual delta.
Conjugate gradient search direction uses the residual delta · case 03
Conjugate gradient search direction uses the residual delta.
Conjugate gradient search direction uses the residual delta · case 04
Conjugate gradient search direction uses the residual delta.
Conjugate gradient search direction uses the residual delta · case 05
Conjugate gradient search direction uses the residual delta.
Trust region accepts steps with a nondecreasing model · case 01
Trust region accepts steps with a nondecreasing model.
Trust region accepts steps with a nondecreasing model · case 02
Trust region accepts steps with a nondecreasing model.
Trust region accepts steps with a nondecreasing model · case 03
Trust region accepts steps with a nondecreasing model.
Trust region accepts steps with a nondecreasing model · case 04
Trust region accepts steps with a nondecreasing model.
Trust region accepts steps with a nondecreasing model · case 05
Trust region accepts steps with a nondecreasing model.
Constraint violation is traded away by objective weighting · case 01
Constraint violation is traded away by objective weighting.
Constraint violation is traded away by objective weighting · case 02
Constraint violation is traded away by objective weighting.
Constraint violation is traded away by objective weighting · case 03
Constraint violation is traded away by objective weighting.
Constraint violation is traded away by objective weighting · case 04
Constraint violation is traded away by objective weighting.
Constraint violation is traded away by objective weighting · case 05
Constraint violation is traded away by objective weighting.
Coordinate descent sweep reads stale coordinates · case 01
Coordinate descent sweep reads stale coordinates.
Coordinate descent sweep reads stale coordinates · case 02
Coordinate descent sweep reads stale coordinates.
Coordinate descent sweep reads stale coordinates · case 03
Coordinate descent sweep reads stale coordinates.
Coordinate descent sweep reads stale coordinates · case 04
Coordinate descent sweep reads stale coordinates.
Coordinate descent sweep reads stale coordinates · case 05
Coordinate descent sweep reads stale coordinates.
Iteration exhaustion is reported as convergence · case 01
Iteration exhaustion is reported as convergence.
Iteration exhaustion is reported as convergence · case 02
Iteration exhaustion is reported as convergence.
Iteration exhaustion is reported as convergence · case 03
Iteration exhaustion is reported as convergence.
Iteration exhaustion is reported as convergence · case 04
Iteration exhaustion is reported as convergence.
Iteration exhaustion is reported as convergence · case 05
Iteration exhaustion is reported as convergence.
Arrival time makes an old measurement look fresh · case 01
Delayed sensor samples enter the current fusion window.
Arrival time makes an old measurement look fresh · case 02
Delayed sensor samples enter the current fusion window.
Arrival time makes an old measurement look fresh · case 03
Delayed sensor samples enter the current fusion window.
Arrival time makes an old measurement look fresh · case 04
Delayed sensor samples enter the current fusion window.
Arrival time makes an old measurement look fresh · case 05
Delayed sensor samples enter the current fusion window.
Sensor clock offset is applied with the wrong sign · case 01
Cross-sensor association rejects simultaneous samples.
Sensor clock offset is applied with the wrong sign · case 02
Cross-sensor association rejects simultaneous samples.
Sensor clock offset is applied with the wrong sign · case 03
Cross-sensor association rejects simultaneous samples.
Sensor clock offset is applied with the wrong sign · case 04
Cross-sensor association rejects simultaneous samples.
Sensor clock offset is applied with the wrong sign · case 05
Cross-sensor association rejects simultaneous samples.
Repeated sensor packet is counted as independent evidence · case 01
Retransmission artificially increases measurement information.
Repeated sensor packet is counted as independent evidence · case 02
Retransmission artificially increases measurement information.
Repeated sensor packet is counted as independent evidence · case 03
Retransmission artificially increases measurement information.
Repeated sensor packet is counted as independent evidence · case 04
Retransmission artificially increases measurement information.
Repeated sensor packet is counted as independent evidence · case 05
Retransmission artificially increases measurement information.
Correlated estimates lose shared uncertainty during fusion · case 01
Reported fused uncertainty is smaller than the shared sensor noise.
Correlated estimates lose shared uncertainty during fusion · case 02
Reported fused uncertainty is smaller than the shared sensor noise.
Correlated estimates lose shared uncertainty during fusion · case 03
Reported fused uncertainty is smaller than the shared sensor noise.
Correlated estimates lose shared uncertainty during fusion · case 04
Reported fused uncertainty is smaller than the shared sensor noise.
Correlated estimates lose shared uncertainty during fusion · case 05
Reported fused uncertainty is smaller than the shared sensor noise.
Missing measurement axis resets a tracked component · case 01
A sensor observing only one axis overwrites the other estimate.
Missing measurement axis resets a tracked component · case 02
A sensor observing only one axis overwrites the other estimate.
Missing measurement axis resets a tracked component · case 03
A sensor observing only one axis overwrites the other estimate.
Missing measurement axis resets a tracked component · case 04
A sensor observing only one axis overwrites the other estimate.
Missing measurement axis resets a tracked component · case 05
A sensor observing only one axis overwrites the other estimate.
Measurement channel permutation leaves covariance behind · case 01
An uncertainty belongs to the wrong sensor channel.
Measurement channel permutation leaves covariance behind · case 02
An uncertainty belongs to the wrong sensor channel.
Measurement channel permutation leaves covariance behind · case 03
An uncertainty belongs to the wrong sensor channel.
Measurement channel permutation leaves covariance behind · case 04
An uncertainty belongs to the wrong sensor channel.
Measurement channel permutation leaves covariance behind · case 05
An uncertainty belongs to the wrong sensor channel.
Innovation gate ignores uncertainty in the predicted state · case 01
Plausible measurements are rejected when prediction uncertainty grows.
Innovation gate ignores uncertainty in the predicted state · case 02
Plausible measurements are rejected when prediction uncertainty grows.
Innovation gate ignores uncertainty in the predicted state · case 03
Plausible measurements are rejected when prediction uncertainty grows.
Innovation gate ignores uncertainty in the predicted state · case 04
Plausible measurements are rejected when prediction uncertainty grows.
Innovation gate ignores uncertainty in the predicted state · case 05
Plausible measurements are rejected when prediction uncertainty grows.
Dropped samples do not accumulate process uncertainty · case 01
Prediction confidence remains too high after a long acquisition gap.
Dropped samples do not accumulate process uncertainty · case 02
Prediction confidence remains too high after a long acquisition gap.
Dropped samples do not accumulate process uncertainty · case 03
Prediction confidence remains too high after a long acquisition gap.
Dropped samples do not accumulate process uncertainty · case 04
Prediction confidence remains too high after a long acquisition gap.
Dropped samples do not accumulate process uncertainty · case 05
Prediction confidence remains too high after a long acquisition gap.
Delayed measurement is applied at the current state epoch · case 01
A late position fix pulls the present estimate backwards along the trajectory.
Delayed measurement is applied at the current state epoch · case 02
A late position fix pulls the present estimate backwards along the trajectory.
Delayed measurement is applied at the current state epoch · case 03
A late position fix pulls the present estimate backwards along the trajectory.
Delayed measurement is applied at the current state epoch · case 04
A late position fix pulls the present estimate backwards along the trajectory.
Delayed measurement is applied at the current state epoch · case 05
A late position fix pulls the present estimate backwards along the trajectory.
Calibration scales measurements but leaves their variances in raw units · case 01
Fusion weights change when an equivalent sensor representation changes units.
Calibration scales measurements but leaves their variances in raw units · case 02
Fusion weights change when an equivalent sensor representation changes units.
Calibration scales measurements but leaves their variances in raw units · case 03
Fusion weights change when an equivalent sensor representation changes units.
Calibration scales measurements but leaves their variances in raw units · case 04
Fusion weights change when an equivalent sensor representation changes units.
Calibration scales measurements but leaves their variances in raw units · case 05
Fusion weights change when an equivalent sensor representation changes units.
Optical camera coordinates are mistaken for forward-left-up coordinates · case 01
Optical camera coordinates are mistaken for forward-left-up coordinates.
Optical camera coordinates are mistaken for forward-left-up coordinates · case 02
Optical camera coordinates are mistaken for forward-left-up coordinates.
Optical camera coordinates are mistaken for forward-left-up coordinates · case 03
Optical camera coordinates are mistaken for forward-left-up coordinates.
Optical camera coordinates are mistaken for forward-left-up coordinates · case 04
Optical camera coordinates are mistaken for forward-left-up coordinates.
Optical camera coordinates are mistaken for forward-left-up coordinates · case 05
Optical camera coordinates are mistaken for forward-left-up coordinates.
North-east-down positions are labeled as east-north-up · case 01
North-east-down positions are labeled as east-north-up.
North-east-down positions are labeled as east-north-up · case 02
North-east-down positions are labeled as east-north-up.
North-east-down positions are labeled as east-north-up · case 03
North-east-down positions are labeled as east-north-up.
North-east-down positions are labeled as east-north-up · case 04
North-east-down positions are labeled as east-north-up.
North-east-down positions are labeled as east-north-up · case 05
North-east-down positions are labeled as east-north-up.
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 ↗