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
State vector from classical elements: Perifocal radial velocity has the wrong sign · case 01
The spacecraft flies its orbit backwards in the radial direction.
State vector from classical elements: Perifocal radial velocity has the wrong sign · case 02
The spacecraft flies its orbit backwards in the radial direction.
State vector from classical elements: Perifocal radial velocity has the wrong sign · case 03
The spacecraft flies its orbit backwards in the radial direction.
State vector from classical elements: Perifocal radial velocity has the wrong sign · case 04
The spacecraft flies its orbit backwards in the radial direction.
State vector from classical elements: Perifocal radial velocity has the wrong sign · case 05
The spacecraft flies its orbit backwards in the radial direction.
State vector from classical elements: Velocity scale uses the current radius · case 01
Speeds are wrong everywhere except on circular orbits.
State vector from classical elements: Velocity scale uses the current radius · case 02
Speeds are wrong everywhere except on circular orbits.
State vector from classical elements: Velocity scale uses the current radius · case 03
Speeds are wrong everywhere except on circular orbits.
State vector from classical elements: Velocity scale uses the current radius · case 04
Speeds are wrong everywhere except on circular orbits.
State vector from classical elements: Velocity scale uses the current radius · case 05
Speeds are wrong everywhere except on circular orbits.
State vector from classical elements: Conic radius uses 1 - e cos nu · case 01
Periapsis and apoapsis distances are swapped.
State vector from classical elements: Conic radius uses 1 - e cos nu · case 02
Periapsis and apoapsis distances are swapped.
State vector from classical elements: Conic radius uses 1 - e cos nu · case 03
Periapsis and apoapsis distances are swapped.
State vector from classical elements: Conic radius uses 1 - e cos nu · case 04
Periapsis and apoapsis distances are swapped.
State vector from classical elements: Conic radius uses 1 - e cos nu · case 05
Periapsis and apoapsis distances are swapped.
State vector from classical elements: True anomaly is left in degrees · case 01
Positions are scattered around the orbit.
State vector from classical elements: True anomaly is left in degrees · case 02
Positions are scattered around the orbit.
State vector from classical elements: True anomaly is left in degrees · case 03
Positions are scattered around the orbit.
State vector from classical elements: True anomaly is left in degrees · case 04
Positions are scattered around the orbit.
State vector from classical elements: True anomaly is left in degrees · case 05
Positions are scattered around the orbit.
Fixed-step RK4 two-body propagation: Second RK4 stage samples a full step ahead · case 01
The integrator degrades to low order and orbits drift.
Fixed-step RK4 two-body propagation: Second RK4 stage samples a full step ahead · case 02
The integrator degrades to low order and orbits drift.
Fixed-step RK4 two-body propagation: Second RK4 stage samples a full step ahead · case 03
The integrator degrades to low order and orbits drift.
Fixed-step RK4 two-body propagation: Second RK4 stage samples a full step ahead · case 04
The integrator degrades to low order and orbits drift.
Fixed-step RK4 two-body propagation: Second RK4 stage samples a full step ahead · case 05
The integrator degrades to low order and orbits drift.
Fixed-step RK4 two-body propagation: RK4 stages are averaged equally · case 01
Energy drifts much faster than expected for the step size.
Fixed-step RK4 two-body propagation: RK4 stages are averaged equally · case 02
Energy drifts much faster than expected for the step size.
Fixed-step RK4 two-body propagation: RK4 stages are averaged equally · case 03
Energy drifts much faster than expected for the step size.
Fixed-step RK4 two-body propagation: RK4 stages are averaged equally · case 04
Energy drifts much faster than expected for the step size.
Fixed-step RK4 two-body propagation: RK4 stages are averaged equally · case 05
Energy drifts much faster than expected for the step size.
Fixed-step RK4 two-body propagation: Acceleration uses |r|^2 in the vector form · case 01
Gravity has the wrong falloff and scales with distance.
Fixed-step RK4 two-body propagation: Acceleration uses |r|^2 in the vector form · case 02
Gravity has the wrong falloff and scales with distance.
Fixed-step RK4 two-body propagation: Acceleration uses |r|^2 in the vector form · case 03
Gravity has the wrong falloff and scales with distance.
Fixed-step RK4 two-body propagation: Acceleration uses |r|^2 in the vector form · case 04
Gravity has the wrong falloff and scales with distance.
Fixed-step RK4 two-body propagation: Acceleration uses |r|^2 in the vector form · case 05
Gravity has the wrong falloff and scales with distance.
Fixed-step RK4 two-body propagation: Third stage reuses the first slope · case 01
The midpoint correction never happens and error grows.
Fixed-step RK4 two-body propagation: Third stage reuses the first slope · case 02
The midpoint correction never happens and error grows.
Fixed-step RK4 two-body propagation: Third stage reuses the first slope · case 03
The midpoint correction never happens and error grows.
Fixed-step RK4 two-body propagation: Third stage reuses the first slope · case 04
The midpoint correction never happens and error grows.
Fixed-step RK4 two-body propagation: Third stage reuses the first slope · case 05
The midpoint correction never happens and error grows.
Fixed-step RK4 two-body propagation: Loop stops one step short of the end time · case 01
The returned state is for T-h rather than T.
Fixed-step RK4 two-body propagation: Loop stops one step short of the end time · case 02
The returned state is for T-h rather than T.
Fixed-step RK4 two-body propagation: Loop stops one step short of the end time · case 03
The returned state is for T-h rather than T.
Fixed-step RK4 two-body propagation: Loop stops one step short of the end time · case 04
The returned state is for T-h rather than T.
Fixed-step RK4 two-body propagation: Loop stops one step short of the end time · case 05
The returned state is for T-h rather than T.
Velocity Verlet two-body propagation: Position drift omits the one-half acceleration factor · case 01
Positions overshoot toward the central body every step.
Velocity Verlet two-body propagation: Position drift omits the one-half acceleration factor · case 02
Positions overshoot toward the central body every step.
Velocity Verlet two-body propagation: Position drift omits the one-half acceleration factor · case 03
Positions overshoot toward the central body every step.
Velocity Verlet two-body propagation: Position drift omits the one-half acceleration factor · case 04
Positions overshoot toward the central body every step.
Velocity Verlet two-body propagation: Position drift omits the one-half acceleration factor · case 05
Positions overshoot toward the central body every step.
Velocity Verlet two-body propagation: Velocity kick uses only the new acceleration · case 01
The scheme becomes first order and energy drifts.
Velocity Verlet two-body propagation: Velocity kick uses only the new acceleration · case 02
The scheme becomes first order and energy drifts.
Velocity Verlet two-body propagation: Velocity kick uses only the new acceleration · case 03
The scheme becomes first order and energy drifts.
Velocity Verlet two-body propagation: Velocity kick uses only the new acceleration · case 04
The scheme becomes first order and energy drifts.
Velocity Verlet two-body propagation: Velocity kick uses only the new acceleration · case 05
The scheme becomes first order and energy drifts.
Velocity Verlet two-body propagation: Acceleration is never carried to the next step · case 01
Every step reuses the initial acceleration.
Velocity Verlet two-body propagation: Acceleration is never carried to the next step · case 02
Every step reuses the initial acceleration.
Velocity Verlet two-body propagation: Acceleration is never carried to the next step · case 03
Every step reuses the initial acceleration.
Velocity Verlet two-body propagation: Acceleration is never carried to the next step · case 04
Every step reuses the initial acceleration.
Velocity Verlet two-body propagation: Acceleration is never carried to the next step · case 05
Every step reuses the initial acceleration.
Velocity Verlet two-body propagation: Final energy uses the initial radius · case 01
Energy diagnostics hide integration drift because the potential never changes.
Velocity Verlet two-body propagation: Final energy uses the initial radius · case 02
Energy diagnostics hide integration drift because the potential never changes.
Velocity Verlet two-body propagation: Final energy uses the initial radius · case 03
Energy diagnostics hide integration drift because the potential never changes.
Velocity Verlet two-body propagation: Final energy uses the initial radius · case 04
Energy diagnostics hide integration drift because the potential never changes.
Velocity Verlet two-body propagation: Final energy uses the initial radius · case 05
Energy diagnostics hide integration drift because the potential never changes.
Velocity Verlet two-body propagation: Initial acceleration is evaluated at the velocity vector · case 01
The first step is kicked in a nonsensical direction.
Velocity Verlet two-body propagation: Initial acceleration is evaluated at the velocity vector · case 02
The first step is kicked in a nonsensical direction.
Velocity Verlet two-body propagation: Initial acceleration is evaluated at the velocity vector · case 03
The first step is kicked in a nonsensical direction.
Velocity Verlet two-body propagation: Initial acceleration is evaluated at the velocity vector · case 04
The first step is kicked in a nonsensical direction.
Velocity Verlet two-body propagation: Initial acceleration is evaluated at the velocity vector · case 05
The first step is kicked in a nonsensical direction.
J2 secular node and perigee drift: J2 rates use the semi-major axis instead of p · case 01
Eccentric orbits get too little nodal regression.
J2 secular node and perigee drift: J2 rates use the semi-major axis instead of p · case 02
Eccentric orbits get too little nodal regression.
J2 secular node and perigee drift: J2 rates use the semi-major axis instead of p · case 03
Eccentric orbits get too little nodal regression.
J2 secular node and perigee drift: J2 rates use the semi-major axis instead of p · case 04
Eccentric orbits get too little nodal regression.
J2 secular node and perigee drift: J2 rates use the semi-major axis instead of p · case 05
Eccentric orbits get too little nodal regression.
J2 secular node and perigee drift: Nodal rate sign ignores retrograde inclinations · case 01
Sun-synchronous retrograde orbits regress instead of precess.
J2 secular node and perigee drift: Nodal rate sign ignores retrograde inclinations · case 02
Sun-synchronous retrograde orbits regress instead of precess.
J2 secular node and perigee drift: Nodal rate sign ignores retrograde inclinations · case 03
Sun-synchronous retrograde orbits regress instead of precess.
J2 secular node and perigee drift: Nodal rate sign ignores retrograde inclinations · case 04
Sun-synchronous retrograde orbits regress instead of precess.
J2 secular node and perigee drift: Nodal rate sign ignores retrograde inclinations · case 05
Sun-synchronous retrograde orbits regress instead of precess.
J2 secular node and perigee drift: Perigee drift factor has the wrong sign · case 01
Critical inclination still shows perigee drift of the wrong direction.
J2 secular node and perigee drift: Perigee drift factor has the wrong sign · case 02
Critical inclination still shows perigee drift of the wrong direction.
J2 secular node and perigee drift: Perigee drift factor has the wrong sign · case 03
Critical inclination still shows perigee drift of the wrong direction.
J2 secular node and perigee drift: Perigee drift factor has the wrong sign · case 04
Critical inclination still shows perigee drift of the wrong direction.
J2 secular node and perigee drift: Perigee drift factor has the wrong sign · case 05
Critical inclination still shows perigee drift of the wrong direction.
J2 secular node and perigee drift: Days are converted to seconds with an hourly factor · case 01
Accumulated drift is 24 times too small.
J2 secular node and perigee drift: Days are converted to seconds with an hourly factor · case 02
Accumulated drift is 24 times too small.
J2 secular node and perigee drift: Days are converted to seconds with an hourly factor · case 03
Accumulated drift is 24 times too small.
J2 secular node and perigee drift: Days are converted to seconds with an hourly factor · case 04
Accumulated drift is 24 times too small.
J2 secular node and perigee drift: Days are converted to seconds with an hourly factor · case 05
Accumulated drift is 24 times too small.
J2 secular node and perigee drift: Sun-synchronous test ignores the precession direction · case 01
Prograde orbits that regress at 0.9856 deg/day are flagged sun-synchronous.
J2 secular node and perigee drift: Sun-synchronous test ignores the precession direction · case 02
Prograde orbits that regress at 0.9856 deg/day are flagged sun-synchronous.
J2 secular node and perigee drift: Sun-synchronous test ignores the precession direction · case 03
Prograde orbits that regress at 0.9856 deg/day are flagged sun-synchronous.
J2 secular node and perigee drift: Sun-synchronous test ignores the precession direction · case 04
Prograde orbits that regress at 0.9856 deg/day are flagged sun-synchronous.
J2 secular node and perigee drift: Sun-synchronous test ignores the precession direction · case 05
Prograde orbits that regress at 0.9856 deg/day are flagged sun-synchronous.
Hyperbolic Kepler solver: Hyperbolic residual has the elliptic sign convention · case 01
Outbound hyperbolic arcs are reported as inbound.
Hyperbolic Kepler solver: Hyperbolic residual has the elliptic sign convention · case 02
Outbound hyperbolic arcs are reported as inbound.
Hyperbolic Kepler solver: Hyperbolic residual has the elliptic sign convention · case 03
Outbound hyperbolic arcs are reported as inbound.
Hyperbolic Kepler solver: Hyperbolic residual has the elliptic sign convention · case 04
Outbound hyperbolic arcs are reported as inbound.
Hyperbolic Kepler solver: Hyperbolic residual has the elliptic sign convention · case 05
Outbound hyperbolic arcs are reported as inbound.
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 ↗