FAILURE MAP
← Case archive

FA-60916 / Bond day-count conventions / Open access

Act/Act ISDA year fraction: century years are all treated as leap years · case 01

Pieces in 1900 or 2100 are divided by 366.

Verified by executionVariant 1 · 8 checks per implementationDownload source bundle ↓JSON ↗

ROOT CAUSE

The leap test only checks divisibility by four.

VERIFIED REPAIR

Apply the full Gregorian rule including the 100 and 400 exceptions.

Unsuccessful approach: Excluding all century years breaks the year 2000, which is a leap year.

Case contract

Inputs start and end [y,m,d]. If end < start return "end before start". Split the interval at each 1 January; each piece contributes its actual days divided by 366 if the piece lies in a Gregorian leap year, else 365. Sum exactly and return the float rounded to 9 decimals.

Why this case matters

Bond accrual and pricing systems depend on exact day-count arithmetic; a single-day error changes settlement cash.

1 / The failure

Exit 1
"""Failure Map reference implementation. Python standard library only."""
import json
import datetime
from fractions import Fraction
N = 1
observations = []
def solve(a, b):
    A = datetime.date(*a)
    B = datetime.date(*b)
    if B < A:
        return 'end before start'
    def leap(y):
        return y % 4 == 0
    total = Fraction(0)
    cur = A
    while cur < B:
        nxt = min(B, datetime.date(cur.year + 1, 1, 1))
        total += Fraction((nxt - cur).days, 366 if leap(cur.year) else 365)
        cur = nxt
    return round(float(total), 9)
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
fixtures = [[['regression gregorian century rule 1', [[1900, 12, 12], [1902, 9, 11]], 1.747945205], ['regression gregorian century rule 2', [[2100, 8, 7], [2101, 12, 23]], 1.378082192], ['partial repair probe 1', [[1998, 2, 12], [2003, 1, 12]], 4.915068493], ['partial repair probe 2', [[1997, 7, 29], [2000, 4, 3]], 2.681495621], ['boundary control 1', [[2024, 1, 1], [2025, 1, 1]], 1.0], ['boundary control 2', [[2024, 3, 1], [2024, 3, 1]], 0.0], ['normal control 1', [[2061, 4, 23], [2062, 11, 8]], 1.545205479], ['normal control 2', [[1965, 4, 28], [1968, 2, 22]], 2.821528558]], [['regression gregorian century rule 1', [[1897, 6, 13], [1901, 8, 19]], 4.183561644], ['regression gregorian century rule 2', [[1900, 4, 25], [1901, 1, 27]], 0.75890411], ['partial repair probe 1', [[2000, 2, 29], [2002, 8, 10]], 2.444277266], ['partial repair probe 2', [[1997, 3, 20], [2001, 12, 29]], 4.778082192], ['boundary control 1', [[2023, 12, 31], [2024, 1, 1]], 0.002739726], ['boundary control 2', [[2024, 1, 1], [2025, 1, 1]], 1.0], ['normal control 1', [[1982, 10, 1], [1983, 8, 2]], 0.835616438], ['normal control 2', [[1918, 6, 30], [1919, 7, 5]], 1.01369863]], [['regression gregorian century rule 1', [[1899, 7, 30], [1900, 6, 17]], 0.882191781], ['regression gregorian century rule 2', [[2100, 12, 19], [2101, 2, 1]], 0.120547945], ['partial repair probe 1', [[1997, 9, 30], [2002, 8, 23]], 4.895890411], ['partial repair probe 2', [[1997, 5, 12], [2000, 7, 5]], 3.149292612], ['boundary control 1', [[2023, 7, 1], [2024, 7, 1]], 1.001377349], ['boundary control 2', [[2023, 12, 31], [2024, 1, 1]], 0.002739726], ['normal control 1', [[2013, 5, 17], [2013, 4, 26]], 'end before start'], ['normal control 2', [[2054, 4, 24], [2054, 11, 1]], 0.523287671]], [['regression gregorian century rule 1', [[1899, 12, 30], [1900, 12, 25]], 0.98630137], ['regression gregorian century rule 2', [[1896, 10, 28], [1900, 5, 19]], 3.55567782], ['partial repair probe 1', [[1997, 4, 13], [2000, 11, 7]], 3.570274721], ['partial repair probe 2', [[2000, 8, 26], [2003, 12, 27]], 3.336028146], ['boundary control 1', [[2023, 12, 31], [2024, 1, 1]], 0.002739726], ['boundary control 2', [[2024, 3, 1], [2024, 3, 1]], 0.0], ['normal control 1', [[1970, 8, 5], [1974, 7, 27]], 3.975342466], ['normal control 2', [[1903, 8, 26], [1909, 7, 28]], 5.920547945]], [['regression gregorian century rule 1', [[1898, 8, 12], [1900, 12, 29]], 2.380821918], ['regression gregorian century rule 2', [[2100, 9, 30], [2106, 3, 21]], 5.471232877], ['partial repair probe 1', [[1998, 2, 27], [2002, 8, 30]], 4.504109589], ['partial repair probe 2', [[1999, 1, 19], [2004, 3, 23]], 5.174728647], ['boundary control 1', [[2024, 3, 1], [2024, 3, 1]], 0.0], ['boundary control 2', [[2023, 12, 31], [2024, 1, 1]], 0.002739726], ['normal control 1', [[1940, 8, 16], [1943, 2, 14]], 2.497597126], ['normal control 2', [[1986, 3, 5], [1986, 11, 19]], 0.709589041]]]
for label, args, expected in fixtures[N-1]:
    check(label, solve(*args), expected)
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)
Boundary fixtureActualExpectedOutcome
regression gregorian century rule 11.7477954941.747945205Failed
regression gregorian century rule 21.376981811.378082192Failed
partial repair probe 14.9150684934.915068493Passed
partial repair probe 22.6814956212.681495621Passed
boundary control 11.01.0Passed
boundary control 20.00.0Passed
normal control 11.5452054791.545205479Passed
normal control 22.8215285582.821528558Passed

SHA-256 / 44fe59aa40a6ce1d306282d528c46e932988c7ffb3d31b2a297ce14d951020b9

2 / The unsuccessful fix

Exit 1
"""Failure Map reference implementation. Python standard library only."""
import json
import datetime
from fractions import Fraction
N = 1
observations = []
def solve(a, b):
    A = datetime.date(*a)
    B = datetime.date(*b)
    if B < A:
        return 'end before start'
    def leap(y):
        return y % 4 == 0 and y % 100 != 0
    total = Fraction(0)
    cur = A
    while cur < B:
        nxt = min(B, datetime.date(cur.year + 1, 1, 1))
        total += Fraction((nxt - cur).days, 366 if leap(cur.year) else 365)
        cur = nxt
    return round(float(total), 9)
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
fixtures = [[['regression gregorian century rule 1', [[1900, 12, 12], [1902, 9, 11]], 1.747945205], ['regression gregorian century rule 2', [[2100, 8, 7], [2101, 12, 23]], 1.378082192], ['partial repair probe 1', [[1998, 2, 12], [2003, 1, 12]], 4.915068493], ['partial repair probe 2', [[1997, 7, 29], [2000, 4, 3]], 2.681495621], ['boundary control 1', [[2024, 1, 1], [2025, 1, 1]], 1.0], ['boundary control 2', [[2024, 3, 1], [2024, 3, 1]], 0.0], ['normal control 1', [[2061, 4, 23], [2062, 11, 8]], 1.545205479], ['normal control 2', [[1965, 4, 28], [1968, 2, 22]], 2.821528558]], [['regression gregorian century rule 1', [[1897, 6, 13], [1901, 8, 19]], 4.183561644], ['regression gregorian century rule 2', [[1900, 4, 25], [1901, 1, 27]], 0.75890411], ['partial repair probe 1', [[2000, 2, 29], [2002, 8, 10]], 2.444277266], ['partial repair probe 2', [[1997, 3, 20], [2001, 12, 29]], 4.778082192], ['boundary control 1', [[2023, 12, 31], [2024, 1, 1]], 0.002739726], ['boundary control 2', [[2024, 1, 1], [2025, 1, 1]], 1.0], ['normal control 1', [[1982, 10, 1], [1983, 8, 2]], 0.835616438], ['normal control 2', [[1918, 6, 30], [1919, 7, 5]], 1.01369863]], [['regression gregorian century rule 1', [[1899, 7, 30], [1900, 6, 17]], 0.882191781], ['regression gregorian century rule 2', [[2100, 12, 19], [2101, 2, 1]], 0.120547945], ['partial repair probe 1', [[1997, 9, 30], [2002, 8, 23]], 4.895890411], ['partial repair probe 2', [[1997, 5, 12], [2000, 7, 5]], 3.149292612], ['boundary control 1', [[2023, 7, 1], [2024, 7, 1]], 1.001377349], ['boundary control 2', [[2023, 12, 31], [2024, 1, 1]], 0.002739726], ['normal control 1', [[2013, 5, 17], [2013, 4, 26]], 'end before start'], ['normal control 2', [[2054, 4, 24], [2054, 11, 1]], 0.523287671]], [['regression gregorian century rule 1', [[1899, 12, 30], [1900, 12, 25]], 0.98630137], ['regression gregorian century rule 2', [[1896, 10, 28], [1900, 5, 19]], 3.55567782], ['partial repair probe 1', [[1997, 4, 13], [2000, 11, 7]], 3.570274721], ['partial repair probe 2', [[2000, 8, 26], [2003, 12, 27]], 3.336028146], ['boundary control 1', [[2023, 12, 31], [2024, 1, 1]], 0.002739726], ['boundary control 2', [[2024, 3, 1], [2024, 3, 1]], 0.0], ['normal control 1', [[1970, 8, 5], [1974, 7, 27]], 3.975342466], ['normal control 2', [[1903, 8, 26], [1909, 7, 28]], 5.920547945]], [['regression gregorian century rule 1', [[1898, 8, 12], [1900, 12, 29]], 2.380821918], ['regression gregorian century rule 2', [[2100, 9, 30], [2106, 3, 21]], 5.471232877], ['partial repair probe 1', [[1998, 2, 27], [2002, 8, 30]], 4.504109589], ['partial repair probe 2', [[1999, 1, 19], [2004, 3, 23]], 5.174728647], ['boundary control 1', [[2024, 3, 1], [2024, 3, 1]], 0.0], ['boundary control 2', [[2023, 12, 31], [2024, 1, 1]], 0.002739726], ['normal control 1', [[1940, 8, 16], [1943, 2, 14]], 2.497597126], ['normal control 2', [[1986, 3, 5], [1986, 11, 19]], 0.709589041]]]
for label, args, expected in fixtures[N-1]:
    check(label, solve(*args), expected)
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)
Boundary fixtureActualExpectedOutcome
regression gregorian century rule 11.7479452051.747945205Passed
regression gregorian century rule 21.3780821921.378082192Passed
partial repair probe 14.9178082194.915068493Failed
partial repair probe 22.6821917812.681495621Failed
boundary control 11.01.0Passed
boundary control 20.00.0Passed
normal control 11.5452054791.545205479Passed
normal control 22.8215285582.821528558Passed

SHA-256 / f7c6c2d814d04cc79ac8ab8ec3cc8747a0e86974c907ae842987975abc7fac36

3 / The verified repair

Exit 0
"""Failure Map reference implementation. Python standard library only."""
import json
import datetime
from fractions import Fraction
N = 1
observations = []
def solve(a, b):
    A = datetime.date(*a)
    B = datetime.date(*b)
    if B < A:
        return 'end before start'
    def leap(y):
        return (y % 4 == 0 and y % 100 != 0) or y % 400 == 0
    total = Fraction(0)
    cur = A
    while cur < B:
        nxt = min(B, datetime.date(cur.year + 1, 1, 1))
        total += Fraction((nxt - cur).days, 366 if leap(cur.year) else 365)
        cur = nxt
    return round(float(total), 9)
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
fixtures = [[['regression gregorian century rule 1', [[1900, 12, 12], [1902, 9, 11]], 1.747945205], ['regression gregorian century rule 2', [[2100, 8, 7], [2101, 12, 23]], 1.378082192], ['partial repair probe 1', [[1998, 2, 12], [2003, 1, 12]], 4.915068493], ['partial repair probe 2', [[1997, 7, 29], [2000, 4, 3]], 2.681495621], ['boundary control 1', [[2024, 1, 1], [2025, 1, 1]], 1.0], ['boundary control 2', [[2024, 3, 1], [2024, 3, 1]], 0.0], ['normal control 1', [[2061, 4, 23], [2062, 11, 8]], 1.545205479], ['normal control 2', [[1965, 4, 28], [1968, 2, 22]], 2.821528558]], [['regression gregorian century rule 1', [[1897, 6, 13], [1901, 8, 19]], 4.183561644], ['regression gregorian century rule 2', [[1900, 4, 25], [1901, 1, 27]], 0.75890411], ['partial repair probe 1', [[2000, 2, 29], [2002, 8, 10]], 2.444277266], ['partial repair probe 2', [[1997, 3, 20], [2001, 12, 29]], 4.778082192], ['boundary control 1', [[2023, 12, 31], [2024, 1, 1]], 0.002739726], ['boundary control 2', [[2024, 1, 1], [2025, 1, 1]], 1.0], ['normal control 1', [[1982, 10, 1], [1983, 8, 2]], 0.835616438], ['normal control 2', [[1918, 6, 30], [1919, 7, 5]], 1.01369863]], [['regression gregorian century rule 1', [[1899, 7, 30], [1900, 6, 17]], 0.882191781], ['regression gregorian century rule 2', [[2100, 12, 19], [2101, 2, 1]], 0.120547945], ['partial repair probe 1', [[1997, 9, 30], [2002, 8, 23]], 4.895890411], ['partial repair probe 2', [[1997, 5, 12], [2000, 7, 5]], 3.149292612], ['boundary control 1', [[2023, 7, 1], [2024, 7, 1]], 1.001377349], ['boundary control 2', [[2023, 12, 31], [2024, 1, 1]], 0.002739726], ['normal control 1', [[2013, 5, 17], [2013, 4, 26]], 'end before start'], ['normal control 2', [[2054, 4, 24], [2054, 11, 1]], 0.523287671]], [['regression gregorian century rule 1', [[1899, 12, 30], [1900, 12, 25]], 0.98630137], ['regression gregorian century rule 2', [[1896, 10, 28], [1900, 5, 19]], 3.55567782], ['partial repair probe 1', [[1997, 4, 13], [2000, 11, 7]], 3.570274721], ['partial repair probe 2', [[2000, 8, 26], [2003, 12, 27]], 3.336028146], ['boundary control 1', [[2023, 12, 31], [2024, 1, 1]], 0.002739726], ['boundary control 2', [[2024, 3, 1], [2024, 3, 1]], 0.0], ['normal control 1', [[1970, 8, 5], [1974, 7, 27]], 3.975342466], ['normal control 2', [[1903, 8, 26], [1909, 7, 28]], 5.920547945]], [['regression gregorian century rule 1', [[1898, 8, 12], [1900, 12, 29]], 2.380821918], ['regression gregorian century rule 2', [[2100, 9, 30], [2106, 3, 21]], 5.471232877], ['partial repair probe 1', [[1998, 2, 27], [2002, 8, 30]], 4.504109589], ['partial repair probe 2', [[1999, 1, 19], [2004, 3, 23]], 5.174728647], ['boundary control 1', [[2024, 3, 1], [2024, 3, 1]], 0.0], ['boundary control 2', [[2023, 12, 31], [2024, 1, 1]], 0.002739726], ['normal control 1', [[1940, 8, 16], [1943, 2, 14]], 2.497597126], ['normal control 2', [[1986, 3, 5], [1986, 11, 19]], 0.709589041]]]
for label, args, expected in fixtures[N-1]:
    check(label, solve(*args), expected)
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)
Boundary fixtureActualExpectedOutcome
regression gregorian century rule 11.7479452051.747945205Passed
regression gregorian century rule 21.3780821921.378082192Passed
partial repair probe 14.9150684934.915068493Passed
partial repair probe 22.6814956212.681495621Passed
boundary control 11.01.0Passed
boundary control 20.00.0Passed
normal control 11.5452054791.545205479Passed
normal control 22.8215285582.821528558Passed

SHA-256 / a055fa60a2be860bec4e38b2cff2f35d3ba2c6cffac5d7e3a52c6e0d52242f08

Verification & scope

A deterministic toy contract stated explicitly in the contract field; no claim of conformance to any published convention text. This reproducer isolates one failure mechanism. Results cover the supplied fixtures. Variants within a family share a test contract and should remain grouped when constructing evaluation splits. Related mechanisms with a shared evaluation_group must also remain together; these controlled models are not independent production incidents.

Observations recorded using Python 3.12.14 at 2026-09-29T14:46:50.030303+00:00.

Case digest / b10e58b7a6dfba41afee3f10c9aaec25531878180ac468c691614dbf625c9e76