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
Hyperbolic Kepler solver: Circular tangent is used on the hyperbolic anomaly · case 01
True anomaly can exceed the asymptote.
Hyperbolic Kepler solver: Circular tangent is used on the hyperbolic anomaly · case 02
True anomaly can exceed the asymptote.
Hyperbolic Kepler solver: Circular tangent is used on the hyperbolic anomaly · case 03
True anomaly can exceed the asymptote.
Hyperbolic Kepler solver: Circular tangent is used on the hyperbolic anomaly · case 04
True anomaly can exceed the asymptote.
Hyperbolic Kepler solver: Circular tangent is used on the hyperbolic anomaly · case 05
True anomaly can exceed the asymptote.
Hyperbolic Kepler solver: Hyperbolic eccentricity ratio is inverted · case 01
True anomalies are compressed toward periapsis.
Hyperbolic Kepler solver: Hyperbolic eccentricity ratio is inverted · case 02
True anomalies are compressed toward periapsis.
Hyperbolic Kepler solver: Hyperbolic eccentricity ratio is inverted · case 03
True anomalies are compressed toward periapsis.
Hyperbolic Kepler solver: Hyperbolic eccentricity ratio is inverted · case 04
True anomalies are compressed toward periapsis.
Hyperbolic Kepler solver: Hyperbolic eccentricity ratio is inverted · case 05
True anomalies are compressed toward periapsis.
Hyperbolic Kepler solver: Asymptote angle uses +1/e · case 01
The reported departure asymptote lies on the wrong side of the latus rectum.
Hyperbolic Kepler solver: Asymptote angle uses +1/e · case 02
The reported departure asymptote lies on the wrong side of the latus rectum.
Hyperbolic Kepler solver: Asymptote angle uses +1/e · case 03
The reported departure asymptote lies on the wrong side of the latus rectum.
Hyperbolic Kepler solver: Asymptote angle uses +1/e · case 04
The reported departure asymptote lies on the wrong side of the latus rectum.
Hyperbolic Kepler solver: Asymptote angle uses +1/e · case 05
The reported departure asymptote lies on the wrong side of the latus rectum.
Stumpff functions for universal variables: Hyperbolic C is negative · case 01
Universal-variable propagation diverges on hyperbolic arcs.
Stumpff functions for universal variables: Hyperbolic C is negative · case 02
Universal-variable propagation diverges on hyperbolic arcs.
Stumpff functions for universal variables: Hyperbolic C is negative · case 03
Universal-variable propagation diverges on hyperbolic arcs.
Stumpff functions for universal variables: Hyperbolic C is negative · case 04
Universal-variable propagation diverges on hyperbolic arcs.
Stumpff functions for universal variables: Hyperbolic C is negative · case 05
Universal-variable propagation diverges on hyperbolic arcs.
Stumpff functions for universal variables: Hyperbolic S uses the elliptic numerator order · case 01
Time of flight on hyperbolic arcs is negative.
Stumpff functions for universal variables: Hyperbolic S uses the elliptic numerator order · case 02
Time of flight on hyperbolic arcs is negative.
Stumpff functions for universal variables: Hyperbolic S uses the elliptic numerator order · case 03
Time of flight on hyperbolic arcs is negative.
Stumpff functions for universal variables: Hyperbolic S uses the elliptic numerator order · case 04
Time of flight on hyperbolic arcs is negative.
Stumpff functions for universal variables: Hyperbolic S uses the elliptic numerator order · case 05
Time of flight on hyperbolic arcs is negative.
Stumpff functions for universal variables: Elliptic S divides by s^2 · case 01
Elliptic time of flight is scaled by sqrt(z).
Stumpff functions for universal variables: Elliptic S divides by s^2 · case 02
Elliptic time of flight is scaled by sqrt(z).
Stumpff functions for universal variables: Elliptic S divides by s^2 · case 03
Elliptic time of flight is scaled by sqrt(z).
Stumpff functions for universal variables: Elliptic S divides by s^2 · case 04
Elliptic time of flight is scaled by sqrt(z).
Stumpff functions for universal variables: Elliptic S divides by s^2 · case 05
Elliptic time of flight is scaled by sqrt(z).
Stumpff functions for universal variables: Near-zero C series has the wrong linear sign · case 01
Near-parabolic orbits get a slightly wrong C.
Stumpff functions for universal variables: Near-zero C series has the wrong linear sign · case 02
Near-parabolic orbits get a slightly wrong C.
Stumpff functions for universal variables: Near-zero C series has the wrong linear sign · case 03
Near-parabolic orbits get a slightly wrong C.
Stumpff functions for universal variables: Near-zero C series has the wrong linear sign · case 04
Near-parabolic orbits get a slightly wrong C.
Stumpff functions for universal variables: Near-zero C series has the wrong linear sign · case 05
Near-parabolic orbits get a slightly wrong C.
Stumpff functions for universal variables: Series branch is used far from zero · case 01
Moderate z values lose accuracy in the truncated series.
Stumpff functions for universal variables: Series branch is used far from zero · case 02
Moderate z values lose accuracy in the truncated series.
Stumpff functions for universal variables: Series branch is used far from zero · case 03
Moderate z values lose accuracy in the truncated series.
Stumpff functions for universal variables: Series branch is used far from zero · case 04
Moderate z values lose accuracy in the truncated series.
Stumpff functions for universal variables: Series branch is used far from zero · case 05
Moderate z values lose accuracy in the truncated series.
Hohmann transfer budget: Transfer semi-major axis is the sum of radii · case 01
Transfer orbit is twice too large.
Hohmann transfer budget: Transfer semi-major axis is the sum of radii · case 02
Transfer orbit is twice too large.
Hohmann transfer budget: Transfer semi-major axis is the sum of radii · case 03
Transfer orbit is twice too large.
Hohmann transfer budget: Transfer semi-major axis is the sum of radii · case 04
Transfer orbit is twice too large.
Hohmann transfer budget: Transfer semi-major axis is the sum of radii · case 05
Transfer orbit is twice too large.
Hohmann transfer budget: Time of flight is a full transfer period · case 01
Arrival burns are scheduled at the departure point.
Hohmann transfer budget: Time of flight is a full transfer period · case 02
Arrival burns are scheduled at the departure point.
Hohmann transfer budget: Time of flight is a full transfer period · case 03
Arrival burns are scheduled at the departure point.
Hohmann transfer budget: Time of flight is a full transfer period · case 04
Arrival burns are scheduled at the departure point.
Hohmann transfer budget: Time of flight is a full transfer period · case 05
Arrival burns are scheduled at the departure point.
Hohmann transfer budget: Arrival burn keeps its sign · case 01
Inward transfers report a negative arrival burn and a reduced total.
Hohmann transfer budget: Arrival burn keeps its sign · case 02
Inward transfers report a negative arrival burn and a reduced total.
Hohmann transfer budget: Arrival burn keeps its sign · case 03
Inward transfers report a negative arrival burn and a reduced total.
Hohmann transfer budget: Arrival burn keeps its sign · case 04
Inward transfers report a negative arrival burn and a reduced total.
Hohmann transfer budget: Arrival burn keeps its sign · case 05
Inward transfers report a negative arrival burn and a reduced total.
Hohmann transfer budget: Arrival speed uses the departure radius · case 01
Arrival burn equals the departure burn geometry.
Hohmann transfer budget: Arrival speed uses the departure radius · case 02
Arrival burn equals the departure burn geometry.
Hohmann transfer budget: Arrival speed uses the departure radius · case 03
Arrival burn equals the departure burn geometry.
Hohmann transfer budget: Arrival speed uses the departure radius · case 04
Arrival burn equals the departure burn geometry.
Hohmann transfer budget: Arrival speed uses the departure radius · case 05
Arrival burn equals the departure burn geometry.
Orbit energy classification: Kinetic energy lacks the one-half factor · case 01
Many bound orbits are classified as escaping.
Orbit energy classification: Kinetic energy lacks the one-half factor · case 02
Many bound orbits are classified as escaping.
Orbit energy classification: Kinetic energy lacks the one-half factor · case 03
Many bound orbits are classified as escaping.
Orbit energy classification: Kinetic energy lacks the one-half factor · case 04
Many bound orbits are classified as escaping.
Orbit energy classification: Kinetic energy lacks the one-half factor · case 05
Many bound orbits are classified as escaping.
Orbit energy classification: Parabolic tolerance is absolute · case 01
The parabolic band depends on the unit system.
Orbit energy classification: Parabolic tolerance is absolute · case 02
The parabolic band depends on the unit system.
Orbit energy classification: Parabolic tolerance is absolute · case 03
The parabolic band depends on the unit system.
Orbit energy classification: Parabolic tolerance is absolute · case 04
The parabolic band depends on the unit system.
Orbit energy classification: Parabolic tolerance is absolute · case 05
The parabolic band depends on the unit system.
Orbit energy classification: Hyperbolic semi-major axis is reported as a magnitude · case 01
Hyperbolic trajectories get positive axes indistinguishable from ellipses.
Orbit energy classification: Hyperbolic semi-major axis is reported as a magnitude · case 02
Hyperbolic trajectories get positive axes indistinguishable from ellipses.
Orbit energy classification: Hyperbolic semi-major axis is reported as a magnitude · case 03
Hyperbolic trajectories get positive axes indistinguishable from ellipses.
Orbit energy classification: Hyperbolic semi-major axis is reported as a magnitude · case 04
Hyperbolic trajectories get positive axes indistinguishable from ellipses.
Orbit energy classification: Hyperbolic semi-major axis is reported as a magnitude · case 05
Hyperbolic trajectories get positive axes indistinguishable from ellipses.
Orbit energy classification: Hyperbolic excess speed omits the factor two · case 01
Departure v-infinity is understated by sqrt(2).
Orbit energy classification: Hyperbolic excess speed omits the factor two · case 02
Departure v-infinity is understated by sqrt(2).
Orbit energy classification: Hyperbolic excess speed omits the factor two · case 03
Departure v-infinity is understated by sqrt(2).
Orbit energy classification: Hyperbolic excess speed omits the factor two · case 04
Departure v-infinity is understated by sqrt(2).
Orbit energy classification: Hyperbolic excess speed omits the factor two · case 05
Departure v-infinity is understated by sqrt(2).
Orbit energy classification: Escape speed is the circular speed · case 01
Escape speed is understated by sqrt(2).
Orbit energy classification: Escape speed is the circular speed · case 02
Escape speed is understated by sqrt(2).
Orbit energy classification: Escape speed is the circular speed · case 03
Escape speed is understated by sqrt(2).
Orbit energy classification: Escape speed is the circular speed · case 04
Escape speed is understated by sqrt(2).
Orbit energy classification: Escape speed is the circular speed · case 05
Escape speed is understated by sqrt(2).
Cylindrical Earth shadow test: Sunward test assumes the sun is along +x · case 01
Satellites are shadowed on the day side when the sun is not on the x axis.
Cylindrical Earth shadow test: Sunward test assumes the sun is along +x · case 02
Satellites are shadowed on the day side when the sun is not on the x axis.
Cylindrical Earth shadow test: Sunward test assumes the sun is along +x · case 03
Satellites are shadowed on the day side when the sun is not on the x axis.
Cylindrical Earth shadow test: Sunward test assumes the sun is along +x · case 04
Satellites are shadowed on the day side when the sun is not on the x axis.
Cylindrical Earth shadow test: Sunward test assumes the sun is along +x · case 05
Satellites are shadowed on the day side when the sun is not on the x axis.
Cylindrical Earth shadow test: Sun direction is not normalized · case 01
Shadow geometry scales with the length of the sun vector.
Cylindrical Earth shadow test: Sun direction is not normalized · case 02
Shadow geometry scales with the length of the sun vector.
Cylindrical Earth shadow test: Sun direction is not normalized · case 03
Shadow geometry scales with the length of the sun vector.
Cylindrical Earth shadow test: Sun direction is not normalized · case 04
Shadow geometry scales with the length of the sun vector.
Cylindrical Earth shadow test: Sun direction is not normalized · case 05
Shadow geometry scales with the length of the sun vector.
Cylindrical Earth shadow test: Shadow test uses the along-axis distance · case 01
Satellites far behind Earth are shadowed even when well off the axis.
Cylindrical Earth shadow test: Shadow test uses the along-axis distance · case 02
Satellites far behind Earth are shadowed even when well off the axis.
Cylindrical Earth shadow test: Shadow test uses the along-axis distance · case 03
Satellites far behind Earth are shadowed even when well off the axis.
Cylindrical Earth shadow test: Shadow test uses the along-axis distance · case 04
Satellites far behind Earth are shadowed even when well off the axis.
Cylindrical Earth shadow test: Shadow test uses the along-axis distance · case 05
Satellites far behind Earth are shadowed even when well off the axis.
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 ↗