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
Age-structured force of infection: susceptibility multiplier · case 01
Groups with reduced susceptibility are infected at full rate.
Age-structured force of infection: susceptibility multiplier · case 02
Groups with reduced susceptibility are infected at full rate.
Age-structured force of infection: susceptibility multiplier · case 03
Groups with reduced susceptibility are infected at full rate.
Age-structured force of infection: susceptibility multiplier · case 04
Groups with reduced susceptibility are infected at full rate.
Age-structured force of infection: susceptibility multiplier · case 05
Groups with reduced susceptibility are infected at full rate.
Age-structured force of infection: susceptible base · case 01
Expected infections include people who are already infected or immune.
Age-structured force of infection: susceptible base · case 02
Expected infections include people who are already infected or immune.
Age-structured force of infection: susceptible base · case 03
Expected infections include people who are already infected or immune.
Age-structured force of infection: susceptible base · case 04
Expected infections include people who are already infected or immune.
Age-structured force of infection: susceptible base · case 05
Expected infections include people who are already infected or immune.
Age-structured force of infection: pressure accumulator reset · case 01
Later age groups accumulate the pressure computed for earlier groups.
Age-structured force of infection: pressure accumulator reset · case 02
Later age groups accumulate the pressure computed for earlier groups.
Age-structured force of infection: pressure accumulator reset · case 03
Later age groups accumulate the pressure computed for earlier groups.
Age-structured force of infection: pressure accumulator reset · case 04
Later age groups accumulate the pressure computed for earlier groups.
Age-structured force of infection: pressure accumulator reset · case 05
Later age groups accumulate the pressure computed for earlier groups.
Windowed renewal-equation Rt estimate: serial-interval lag origin · case 01
Same-day cases are counted as their own infectors.
Windowed renewal-equation Rt estimate: serial-interval lag origin · case 02
Same-day cases are counted as their own infectors.
Windowed renewal-equation Rt estimate: serial-interval lag origin · case 03
Same-day cases are counted as their own infectors.
Windowed renewal-equation Rt estimate: serial-interval lag origin · case 04
Same-day cases are counted as their own infectors.
Windowed renewal-equation Rt estimate: serial-interval lag origin · case 05
Same-day cases are counted as their own infectors.
Windowed renewal-equation Rt estimate: truncated weight renormalisation · case 01
Early-window Rt is biased upward and unnormalised weights distort all estimates.
Windowed renewal-equation Rt estimate: truncated weight renormalisation · case 02
Early-window Rt is biased upward and unnormalised weights distort all estimates.
Windowed renewal-equation Rt estimate: truncated weight renormalisation · case 03
Early-window Rt is biased upward and unnormalised weights distort all estimates.
Windowed renewal-equation Rt estimate: truncated weight renormalisation · case 04
Early-window Rt is biased upward and unnormalised weights distort all estimates.
Windowed renewal-equation Rt estimate: truncated weight renormalisation · case 05
Early-window Rt is biased upward and unnormalised weights distort all estimates.
Windowed renewal-equation Rt estimate: window start · case 01
Rt lags one day behind and excludes the most recent incidence.
Windowed renewal-equation Rt estimate: window start · case 02
Rt lags one day behind and excludes the most recent incidence.
Windowed renewal-equation Rt estimate: window start · case 03
Rt lags one day behind and excludes the most recent incidence.
Windowed renewal-equation Rt estimate: window start · case 04
Rt lags one day behind and excludes the most recent incidence.
Windowed renewal-equation Rt estimate: window start · case 05
Rt lags one day behind and excludes the most recent incidence.
Windowed renewal-equation Rt estimate: window feasibility · case 01
A window touching day 0 is estimated using a day with no infection history.
Windowed renewal-equation Rt estimate: window feasibility · case 02
A window touching day 0 is estimated using a day with no infection history.
Windowed renewal-equation Rt estimate: window feasibility · case 03
A window touching day 0 is estimated using a day with no infection history.
Windowed renewal-equation Rt estimate: window feasibility · case 04
A window touching day 0 is estimated using a day with no infection history.
Windowed renewal-equation Rt estimate: window feasibility · case 05
A window touching day 0 is estimated using a day with no infection history.
Windowed renewal-equation Rt estimate: zero pressure result · case 01
Windows with no prior infections report Rt=0 instead of undefined.
Windowed renewal-equation Rt estimate: zero pressure result · case 02
Windows with no prior infections report Rt=0 instead of undefined.
Windowed renewal-equation Rt estimate: zero pressure result · case 03
Windows with no prior infections report Rt=0 instead of undefined.
Windowed renewal-equation Rt estimate: zero pressure result · case 04
Windows with no prior infections report Rt=0 instead of undefined.
Windowed renewal-equation Rt estimate: zero pressure result · case 05
Windows with no prior infections report Rt=0 instead of undefined.
Epidemic growth rate, doubling time and R: zero-count day alignment · case 01
Growth rates are overstated whenever zero-count days are dropped.
Epidemic growth rate, doubling time and R: zero-count day alignment · case 02
Growth rates are overstated whenever zero-count days are dropped.
Epidemic growth rate, doubling time and R: zero-count day alignment · case 03
Growth rates are overstated whenever zero-count days are dropped.
Epidemic growth rate, doubling time and R: zero-count day alignment · case 04
Growth rates are overstated whenever zero-count days are dropped.
Epidemic growth rate, doubling time and R: zero-count day alignment · case 05
Growth rates are overstated whenever zero-count days are dropped.
Epidemic growth rate, doubling time and R: slope estimator · case 01
Growth rate depends only on the first and last counts and on how many days were positive.
Epidemic growth rate, doubling time and R: slope estimator · case 02
Growth rate depends only on the first and last counts and on how many days were positive.
Epidemic growth rate, doubling time and R: slope estimator · case 03
Growth rate depends only on the first and last counts and on how many days were positive.
Epidemic growth rate, doubling time and R: slope estimator · case 04
Growth rate depends only on the first and last counts and on how many days were positive.
Epidemic growth rate, doubling time and R: slope estimator · case 05
Growth rate depends only on the first and last counts and on how many days were positive.
Epidemic growth rate, doubling time and R: declining doubling time · case 01
Declining epidemics report a negative doubling time.
Epidemic growth rate, doubling time and R: declining doubling time · case 02
Declining epidemics report a negative doubling time.
Epidemic growth rate, doubling time and R: declining doubling time · case 03
Declining epidemics report a negative doubling time.
Epidemic growth rate, doubling time and R: declining doubling time · case 04
Declining epidemics report a negative doubling time.
Epidemic growth rate, doubling time and R: declining doubling time · case 05
Declining epidemics report a negative doubling time.
Epidemic growth rate, doubling time and R: log base · case 01
Growth rates are scaled by a constant factor and R is understated.
Epidemic growth rate, doubling time and R: log base · case 02
Growth rates are scaled by a constant factor and R is understated.
Epidemic growth rate, doubling time and R: log base · case 03
Growth rates are scaled by a constant factor and R is understated.
Epidemic growth rate, doubling time and R: log base · case 04
Growth rates are scaled by a constant factor and R is understated.
Epidemic growth rate, doubling time and R: log base · case 05
Growth rates are scaled by a constant factor and R is understated.
Epidemic growth rate, doubling time and R: stage split · case 01
R is overstated whenever there is a latent period.
Epidemic growth rate, doubling time and R: stage split · case 02
R is overstated whenever there is a latent period.
Epidemic growth rate, doubling time and R: stage split · case 03
R is overstated whenever there is a latent period.
Epidemic growth rate, doubling time and R: stage split · case 04
R is overstated whenever there is a latent period.
Epidemic growth rate, doubling time and R: stage split · case 05
R is overstated whenever there is a latent period.
Epidemic growth rate, doubling time and R: reproduction mapping · case 01
R ignores the two-stage structure of the generation interval.
Epidemic growth rate, doubling time and R: reproduction mapping · case 02
R ignores the two-stage structure of the generation interval.
Epidemic growth rate, doubling time and R: reproduction mapping · case 03
R ignores the two-stage structure of the generation interval.
Epidemic growth rate, doubling time and R: reproduction mapping · case 04
R ignores the two-stage structure of the generation interval.
Epidemic growth rate, doubling time and R: reproduction mapping · case 05
R ignores the two-stage structure of the generation interval.
SIQR testing and isolation model: isolated infectiousness · case 01
Isolation does not slow transmission.
SIQR testing and isolation model: isolated infectiousness · case 02
Isolation does not slow transmission.
SIQR testing and isolation model: isolated infectiousness · case 03
Isolation does not slow transmission.
SIQR testing and isolation model: isolated infectiousness · case 04
Isolation does not slow transmission.
SIQR testing and isolation model: isolated infectiousness · case 05
Isolation does not slow transmission.
SIQR testing and isolation model: competing exits cap · case 01
With aggressive testing prevalence becomes negative.
SIQR testing and isolation model: competing exits cap · case 02
With aggressive testing prevalence becomes negative.
SIQR testing and isolation model: competing exits cap · case 03
With aggressive testing prevalence becomes negative.
SIQR testing and isolation model: competing exits cap · case 04
With aggressive testing prevalence becomes negative.
SIQR testing and isolation model: competing exits cap · case 05
With aggressive testing prevalence becomes negative.
SIQR testing and isolation model: release rate · case 01
Isolated people pile up and are released only in proportion to new detections.
SIQR testing and isolation model: release rate · case 02
Isolated people pile up and are released only in proportion to new detections.
SIQR testing and isolation model: release rate · case 03
Isolated people pile up and are released only in proportion to new detections.
SIQR testing and isolation model: release rate · case 04
Isolated people pile up and are released only in proportion to new detections.
SIQR testing and isolation model: release rate · case 05
Isolated people pile up and are released only in proportion to new detections.
SIQR testing and isolation model: release destination · case 01
People leaving isolation become susceptible again.
SIQR testing and isolation model: release destination · case 02
People leaving isolation become susceptible again.
SIQR testing and isolation model: release destination · case 03
People leaving isolation become susceptible again.
SIQR testing and isolation model: release destination · case 04
People leaving isolation become susceptible again.
SIQR testing and isolation model: release destination · case 05
People leaving isolation become susceptible again.
SIQR testing and isolation model: detection series · case 01
Reported daily detections accumulate instead of showing new positives.
SIQR testing and isolation model: detection series · case 02
Reported daily detections accumulate instead of showing new positives.
SIQR testing and isolation model: detection series · case 03
Reported daily detections accumulate instead of showing new positives.
SIQR testing and isolation model: detection series · case 04
Reported daily detections accumulate instead of showing new positives.
SIQR testing and isolation model: detection series · case 05
Reported daily detections accumulate instead of showing new positives.
Hospital occupancy from delayed admissions: admission delay direction · case 01
Admissions precede the infections that cause them.
Hospital occupancy from delayed admissions: admission delay direction · case 02
Admissions precede the infections that cause them.
Hospital occupancy from delayed admissions: admission delay direction · case 03
Admissions precede the infections that cause them.
Hospital occupancy from delayed admissions: admission delay direction · case 04
Admissions precede the infections that cause them.
Hospital occupancy from delayed admissions: admission delay direction · case 05
Admissions precede the infections that cause them.
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 ↗