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
The 1k to 1d step is counted as two ranks · case 01
Games across the kyu/dan boundary get an extra handicap stone.
The 1k to 1d step is counted as two ranks · case 02
Games across the kyu/dan boundary get an extra handicap stone.
The 1k to 1d step is counted as two ranks · case 03
Games across the kyu/dan boundary get an extra handicap stone.
The 1k to 1d step is counted as two ranks · case 04
Games across the kyu/dan boundary get an extra handicap stone.
The 1k to 1d step is counted as two ranks · case 05
Games across the kyu/dan boundary get an extra handicap stone.
The stronger player is given black · case 01
Handicap stones go to the stronger player.
The stronger player is given black · case 02
Handicap stones go to the stronger player.
The stronger player is given black · case 03
Handicap stones go to the stronger player.
The stronger player is given black · case 04
Handicap stones go to the stronger player.
The stronger player is given black · case 05
Handicap stones go to the stronger player.
Handicap exceeds nine stones · case 01
Large rank gaps produce impossible handicaps.
Handicap exceeds nine stones · case 02
Large rank gaps produce impossible handicaps.
Handicap exceeds nine stones · case 03
Large rank gaps produce impossible handicaps.
Handicap exceeds nine stones · case 04
Large rank gaps produce impossible handicaps.
Handicap exceeds nine stones · case 05
Large rank gaps produce impossible handicaps.
A one-rank gap is given a one-stone handicap · case 01
Adjacent ranks receive a stone instead of a no-komi game.
A one-rank gap is given a one-stone handicap · case 02
Adjacent ranks receive a stone instead of a no-komi game.
A one-rank gap is given a one-stone handicap · case 03
Adjacent ranks receive a stone instead of a no-komi game.
A one-rank gap is given a one-stone handicap · case 04
Adjacent ranks receive a stone instead of a no-komi game.
A one-rank gap is given a one-stone handicap · case 05
Adjacent ranks receive a stone instead of a no-komi game.
Out-of-range kyu ranks are accepted · case 01
0k and 31k produce handicaps.
Out-of-range kyu ranks are accepted · case 02
0k and 31k produce handicaps.
Out-of-range kyu ranks are accepted · case 03
0k and 31k produce handicaps.
Out-of-range kyu ranks are accepted · case 04
0k and 31k produce handicaps.
Out-of-range kyu ranks are accepted · case 05
0k and 31k produce handicaps.
Non-consecutive passes end the game · case 01
A pass, a move, then an opponent pass is treated as two consecutive passes.
Non-consecutive passes end the game · case 02
A pass, a move, then an opponent pass is treated as two consecutive passes.
Non-consecutive passes end the game · case 03
A pass, a move, then an opponent pass is treated as two consecutive passes.
Non-consecutive passes end the game · case 04
A pass, a move, then an opponent pass is treated as two consecutive passes.
Non-consecutive passes end the game · case 05
A pass, a move, then an opponent pass is treated as two consecutive passes.
Two passes by the same player end the game · case 01
A player passing twice in a row (e.g. during free placement) ends the game.
Two passes by the same player end the game · case 02
A player passing twice in a row (e.g. during free placement) ends the game.
Two passes by the same player end the game · case 03
A player passing twice in a row (e.g. during free placement) ends the game.
Two passes by the same player end the game · case 04
A player passing twice in a row (e.g. during free placement) ends the game.
Two passes by the same player end the game · case 05
A player passing twice in a row (e.g. during free placement) ends the game.
The resigning player is recorded as winner · case 01
Resignations invert the result.
The resigning player is recorded as winner · case 02
Resignations invert the result.
The resigning player is recorded as winner · case 03
Resignations invert the result.
The resigning player is recorded as winner · case 04
Resignations invert the result.
The resigning player is recorded as winner · case 05
Resignations invert the result.
Timeouts are treated as ordinary moves · case 01
A game lost on time continues and may later end by passes.
Timeouts are treated as ordinary moves · case 02
A game lost on time continues and may later end by passes.
Timeouts are treated as ordinary moves · case 03
A game lost on time continues and may later end by passes.
Timeouts are treated as ordinary moves · case 04
A game lost on time continues and may later end by passes.
Timeouts are treated as ordinary moves · case 05
A game lost on time continues and may later end by passes.
Captured stones are credited to the captured side · case 01
Replay prisoner totals are swapped for every capture.
Captured stones are credited to the captured side · case 02
Replay prisoner totals are swapped for every capture.
Captured stones are credited to the captured side · case 03
Replay prisoner totals are swapped for every capture.
Captured stones are credited to the captured side · case 04
Replay prisoner totals are swapped for every capture.
Captured stones are credited to the captured side · case 05
Replay prisoner totals are swapped for every capture.
Replay stops at the first pass · case 01
Moves after an early pass are ignored.
Replay stops at the first pass · case 02
Moves after an early pass are ignored.
Replay stops at the first pass · case 03
Moves after an early pass are ignored.
Replay stops at the first pass · case 04
Moves after an early pass are ignored.
Replay stops at the first pass · case 05
Moves after an early pass are ignored.
Suicidal moves are replayed as legal · case 01
A libertyless chain remains on the replay board.
Suicidal moves are replayed as legal · case 02
A libertyless chain remains on the replay board.
Suicidal moves are replayed as legal · case 03
A libertyless chain remains on the replay board.
Suicidal moves are replayed as legal · case 04
A libertyless chain remains on the replay board.
Suicidal moves are replayed as legal · case 05
A libertyless chain remains on the replay board.
Points exactly at the reach distance are left neutral · case 01
The estimate shrinks every moyo by one ring.
Points exactly at the reach distance are left neutral · case 02
The estimate shrinks every moyo by one ring.
Points exactly at the reach distance are left neutral · case 03
The estimate shrinks every moyo by one ring.
Points exactly at the reach distance are left neutral · case 04
The estimate shrinks every moyo by one ring.
Points exactly at the reach distance are left neutral · case 05
The estimate shrinks every moyo by one ring.
Equidistant black and white stones give the point to one side · case 01
Boundary points are assigned by scan order.
Equidistant black and white stones give the point to one side · case 02
Boundary points are assigned by scan order.
Equidistant black and white stones give the point to one side · case 03
Boundary points are assigned by scan order.
Equidistant black and white stones give the point to one side · case 04
Boundary points are assigned by scan order.
Equidistant black and white stones give the point to one side · case 05
Boundary points are assigned by scan order.
Diagonal stones count as distance one · case 01
Estimated territory extends diagonally past the intended shape.
Diagonal stones count as distance one · case 02
Estimated territory extends diagonally past the intended shape.
Diagonal stones count as distance one · case 03
Estimated territory extends diagonally past the intended shape.
Diagonal stones count as distance one · case 04
Estimated territory extends diagonally past the intended shape.
Diagonal stones count as distance one · case 05
Estimated territory extends diagonally past the intended shape.
Prisoners are filled into their captor territory · case 01
Each side loses territory for stones it captured.
Prisoners are filled into their captor territory · case 02
Each side loses territory for stones it captured.
Prisoners are filled into their captor territory · case 03
Each side loses territory for stones it captured.
Prisoners are filled into their captor territory · case 04
Each side loses territory for stones it captured.
Prisoners are filled into their captor territory · case 05
Each side loses territory for stones it captured.
Komi tenths are multiplied again by ten · case 01
A 6.5 komi is applied as 65 points.
Komi tenths are multiplied again by ten · case 02
A 6.5 komi is applied as 65 points.
Komi tenths are multiplied again by ten · case 03
A 6.5 komi is applied as 65 points.
Komi tenths are multiplied again by ten · case 04
A 6.5 komi is applied as 65 points.
Komi tenths are multiplied again by ten · case 05
A 6.5 komi is applied as 65 points.
Proposals listing the same stones in a different order disagree · case 01
Players who marked the same stones in a different order are sent back to dispute.
Proposals listing the same stones in a different order disagree · case 02
Players who marked the same stones in a different order are sent back to dispute.
Proposals listing the same stones in a different order disagree · case 03
Players who marked the same stones in a different order are sent back to dispute.
Proposals listing the same stones in a different order disagree · case 04
Players who marked the same stones in a different order are sent back to dispute.
Proposals listing the same stones in a different order disagree · case 05
Players who marked the same stones in a different order are sent back to dispute.
Only stones proposed by black are reported as disputed · case 01
Stones only white marked are omitted from the dispute list.
Only stones proposed by black are reported as disputed · case 02
Stones only white marked are omitted from the dispute list.
Only stones proposed by black are reported as disputed · case 03
Stones only white marked are omitted from the dispute list.
Only stones proposed by black are reported as disputed · case 04
Stones only white marked are omitted from the dispute list.
Only stones proposed by black are reported as disputed · case 05
Stones only white marked are omitted from the dispute list.
The agreed dead list is the raw black proposal · case 01
Agreed results carry duplicates and caller order.
The agreed dead list is the raw black proposal · case 02
Agreed results carry duplicates and caller order.
The agreed dead list is the raw black proposal · case 03
Agreed results carry duplicates and caller order.
The agreed dead list is the raw black proposal · case 04
Agreed results carry duplicates and caller order.
The agreed dead list is the raw black proposal · case 05
Agreed results carry duplicates and caller order.
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 ↗