FA-62186 / Currency rounding and FX conversion / Open access

The disclosed markup is measured against the offered rate · case 01

A 3% markup is disclosed as 2.92%.

Verified by executionVariant 1 · 6 checks per implementationDownload source bundle ↓JSON ↗

ROOT CAUSE

The disclosure divides the rate difference by the offered rate instead of the reference rate.

VERIFIED REPAIR

Disclose (offered/ref - 1)*100 using the rounded offered rate.

Unsuccessful approach: Disclosing bps/100 ignores the effect of rounding the offered rate.

Case contract

solve(amount, ref_rate, markup_bps, card_exp): amount in merchant currency, ref_rate cardholder units per merchant unit, markup_bps an integer (negative for a promotional discount). The offered rate is ref*(1+bps/10000) rounded half-up to 6 decimals, and the card amount is amount*offered rounded half-up at card_exp. The disclosed markup percentage is (offered/ref - 1)*100 computed from the rounded offered rate and rounded toward positive infinity to 2 decimals so it never understates. Return [offered, card_amount, markup_pct] as strings.

Why this case matters

Currency amounts must be rounded at the right stage and in the right unit, or ledgers, quotes and settlements drift by minor units.

1 / The failure

Exit 1

"""Failure Map reference implementation. Python standard library only."""
import json
from decimal import Decimal, ROUND_HALF_EVEN, ROUND_HALF_UP, ROUND_FLOOR, ROUND_CEILING, ROUND_DOWN, ROUND_UP, ROUND_HALF_DOWN
N = 1
observations = []
def solve(amount, ref_rate, markup_bps, card_exp):
    ref = Decimal(ref_rate)
    raw = ref * (1 + Decimal(markup_bps) / 10000)
    dcc = raw.quantize(Decimal('0.000001'), rounding=ROUND_HALF_UP)
    card = (Decimal(amount) * dcc).quantize(Decimal(1).scaleb(-card_exp), rounding=ROUND_HALF_UP)
    pct = ((dcc - ref) / dcc * 100).quantize(Decimal('0.01'), rounding=ROUND_CEILING)
    return [format(dcc, 'f'), format(card, 'f'), format(pct, 'f')]
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
cases = [[('regression markup-disclosure-base 1', ('100.00', '1.083471', 300, 2), ['1.115975', '111.60', '3.00']),
  ('regression markup-disclosure-base 2', ('5000', '0.00662', 350, 2), ['0.006852', '34.26', '3.51']),
  ('partial repair guard 2', ('49.99', '151.3742', 275, 0), ['155.536991', '7775', '2.76']),
  ('control: dinar card', ('250.00', '0.307133', 400, 3), ['0.319418', '79.855', '4.00']),
  ('control: promotional discount', ('80.00', '0.913377', -50, 2), ['0.908810', '72.70', '-0.50']),
  ('control: zero markup', ('12.34', '1.2', 0, 2), ['1.200000', '14.81', '0.00'])],
 [('regression markup-disclosure-base 1', ('49.99', '151.3742', 275, 0), ['155.536991', '7775', '2.76']),
  ('regression markup-disclosure-base 2', ('250.00', '0.307133', 400, 3), ['0.319418', '79.855', '4.00']),
  ('partial repair guard 1', ('3353.43', '37.36650', 275, 0), ['38.394079', '128752', '2.76']),
  ('partial repair guard 2', ('1171.26', '0.4683115', -50, 3), ['0.465970', '545.772', '-0.49']),
  ('control: promotional discount', ('80.00', '0.913377', -50, 2), ['0.908810', '72.70', '-0.50']),
  ('control: zero markup', ('12.34', '1.2', 0, 2), ['1.200000', '14.81', '0.00']),
  ('control: euro to dollar', ('100.00', '1.083471', 300, 2), ['1.115975', '111.60', '3.00']),
  ('control: yen to dollar', ('5000', '0.00662', 350, 2), ['0.006852', '34.26', '3.51'])],
 [('regression markup-disclosure-base 1', ('3353.43', '37.36650', 275, 0), ['38.394079', '128752', '2.76']),
  ('regression markup-disclosure-base 2', ('1171.26', '0.4683115', -50, 3), ['0.465970', '545.772', '-0.49']),
  ('partial repair guard 1', ('157.22', '31.44890', 150, 3), ['31.920634', '5018.562', '1.51']),
  ('partial repair guard 2', ('3157.31', '0.0891399', 150, 3), ['0.090477', '285.664', '1.51']),
  ('control: euro to dollar', ('100.00', '1.083471', 300, 2), ['1.115975', '111.60', '3.00']),
  ('control: yen to dollar', ('5000', '0.00662', 350, 2), ['0.006852', '34.26', '3.51']),
  ('control: to yen card', ('49.99', '151.3742', 275, 0), ['155.536991', '7775', '2.76']),
  ('control: dinar card', ('250.00', '0.307133', 400, 3), ['0.319418', '79.855', '4.00'])],
 [('regression markup-disclosure-base 1', ('3884.23', '183.00464', 350, 0), ['189.409802', '735711', '3.50']),
  ('regression markup-disclosure-base 2', ('157.22', '31.44890', 150, 3), ['31.920634', '5018.562', '1.51']),
  ('partial repair guard 1', ('508.07', '12.046349', 275, 0), ['12.377624', '6289', '2.76']),
  ('partial repair guard 2', ('3826.89', '1.5416740', -50, 2), ['1.533966', '5870.32', '-0.49']),
  ('control: dinar card', ('250.00', '0.307133', 400, 3), ['0.319418', '79.855', '4.00']),
  ('control: promotional discount', ('80.00', '0.913377', -50, 2), ['0.908810', '72.70', '-0.50']),
  ('control: zero markup', ('12.34', '1.2', 0, 2), ['1.200000', '14.81', '0.00']),
  ('control: euro to dollar', ('100.00', '1.083471', 300, 2), ['1.115975', '111.60', '3.00'])],
 [('regression markup-disclosure-base 1', ('3157.31', '0.0891399', 150, 3), ['0.090477', '285.664', '1.51']),
  ('regression markup-disclosure-base 2', ('3408.26', '18.286201', 150, 2), ['18.560494', '63258.99', '1.50']),
  ('partial repair guard 1', ('419.33', '78.12165', 350, 2), ['80.855908', '33905.31', '3.51']),
  ('partial repair guard 2', ('2378.17', '122.42736', 300, 2), ['126.100181', '299887.67', '3.01']),
  ('control: euro to dollar', ('100.00', '1.083471', 300, 2), ['1.115975', '111.60', '3.00']),
  ('control: yen to dollar', ('5000', '0.00662', 350, 2), ['0.006852', '34.26', '3.51']),
  ('control: to yen card', ('49.99', '151.3742', 275, 0), ['155.536991', '7775', '2.76']),
  ('control: dinar card', ('250.00', '0.307133', 400, 3), ['0.319418', '79.855', '4.00'])]]
for label, args, expected in cases[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 fixture	Actual	Expected	Outcome
regression markup-disclosure-base 1	`['1.115975', '111.60', '2.92']`	`['1.115975', '111.60', '3.00']`	Failed
regression markup-disclosure-base 2	`['0.006852', '34.26', '3.39']`	`['0.006852', '34.26', '3.51']`	Failed
partial repair guard 2	`['155.536991', '7775', '2.68']`	`['155.536991', '7775', '2.76']`	Failed
control: dinar card	`['0.319418', '79.855', '3.85']`	`['0.319418', '79.855', '4.00']`	Failed
control: promotional discount	`['0.908810', '72.70', '-0.50']`	`['0.908810', '72.70', '-0.50']`	Passed
control: zero markup	`['1.200000', '14.81', '0.00']`	`['1.200000', '14.81', '0.00']`	Passed

SHA-256 / 1a60cb4e8bddc04dc932513915314dd84d8b0300629845554c41287522d4ddb5

2 / The unsuccessful fix

Exit 1

"""Failure Map reference implementation. Python standard library only."""
import json
from decimal import Decimal, ROUND_HALF_EVEN, ROUND_HALF_UP, ROUND_FLOOR, ROUND_CEILING, ROUND_DOWN, ROUND_UP, ROUND_HALF_DOWN
N = 1
observations = []
def solve(amount, ref_rate, markup_bps, card_exp):
    ref = Decimal(ref_rate)
    raw = ref * (1 + Decimal(markup_bps) / 10000)
    dcc = raw.quantize(Decimal('0.000001'), rounding=ROUND_HALF_UP)
    card = (Decimal(amount) * dcc).quantize(Decimal(1).scaleb(-card_exp), rounding=ROUND_HALF_UP)
    pct = (Decimal(markup_bps) / 100).quantize(Decimal('0.01'), rounding=ROUND_CEILING)
    return [format(dcc, 'f'), format(card, 'f'), format(pct, 'f')]
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
cases = [[('regression markup-disclosure-base 1', ('100.00', '1.083471', 300, 2), ['1.115975', '111.60', '3.00']),
  ('regression markup-disclosure-base 2', ('5000', '0.00662', 350, 2), ['0.006852', '34.26', '3.51']),
  ('partial repair guard 2', ('49.99', '151.3742', 275, 0), ['155.536991', '7775', '2.76']),
  ('control: dinar card', ('250.00', '0.307133', 400, 3), ['0.319418', '79.855', '4.00']),
  ('control: promotional discount', ('80.00', '0.913377', -50, 2), ['0.908810', '72.70', '-0.50']),
  ('control: zero markup', ('12.34', '1.2', 0, 2), ['1.200000', '14.81', '0.00'])],
 [('regression markup-disclosure-base 1', ('49.99', '151.3742', 275, 0), ['155.536991', '7775', '2.76']),
  ('regression markup-disclosure-base 2', ('250.00', '0.307133', 400, 3), ['0.319418', '79.855', '4.00']),
  ('partial repair guard 1', ('3353.43', '37.36650', 275, 0), ['38.394079', '128752', '2.76']),
  ('partial repair guard 2', ('1171.26', '0.4683115', -50, 3), ['0.465970', '545.772', '-0.49']),
  ('control: promotional discount', ('80.00', '0.913377', -50, 2), ['0.908810', '72.70', '-0.50']),
  ('control: zero markup', ('12.34', '1.2', 0, 2), ['1.200000', '14.81', '0.00']),
  ('control: euro to dollar', ('100.00', '1.083471', 300, 2), ['1.115975', '111.60', '3.00']),
  ('control: yen to dollar', ('5000', '0.00662', 350, 2), ['0.006852', '34.26', '3.51'])],
 [('regression markup-disclosure-base 1', ('3353.43', '37.36650', 275, 0), ['38.394079', '128752', '2.76']),
  ('regression markup-disclosure-base 2', ('1171.26', '0.4683115', -50, 3), ['0.465970', '545.772', '-0.49']),
  ('partial repair guard 1', ('157.22', '31.44890', 150, 3), ['31.920634', '5018.562', '1.51']),
  ('partial repair guard 2', ('3157.31', '0.0891399', 150, 3), ['0.090477', '285.664', '1.51']),
  ('control: euro to dollar', ('100.00', '1.083471', 300, 2), ['1.115975', '111.60', '3.00']),
  ('control: yen to dollar', ('5000', '0.00662', 350, 2), ['0.006852', '34.26', '3.51']),
  ('control: to yen card', ('49.99', '151.3742', 275, 0), ['155.536991', '7775', '2.76']),
  ('control: dinar card', ('250.00', '0.307133', 400, 3), ['0.319418', '79.855', '4.00'])],
 [('regression markup-disclosure-base 1', ('3884.23', '183.00464', 350, 0), ['189.409802', '735711', '3.50']),
  ('regression markup-disclosure-base 2', ('157.22', '31.44890', 150, 3), ['31.920634', '5018.562', '1.51']),
  ('partial repair guard 1', ('508.07', '12.046349', 275, 0), ['12.377624', '6289', '2.76']),
  ('partial repair guard 2', ('3826.89', '1.5416740', -50, 2), ['1.533966', '5870.32', '-0.49']),
  ('control: dinar card', ('250.00', '0.307133', 400, 3), ['0.319418', '79.855', '4.00']),
  ('control: promotional discount', ('80.00', '0.913377', -50, 2), ['0.908810', '72.70', '-0.50']),
  ('control: zero markup', ('12.34', '1.2', 0, 2), ['1.200000', '14.81', '0.00']),
  ('control: euro to dollar', ('100.00', '1.083471', 300, 2), ['1.115975', '111.60', '3.00'])],
 [('regression markup-disclosure-base 1', ('3157.31', '0.0891399', 150, 3), ['0.090477', '285.664', '1.51']),
  ('regression markup-disclosure-base 2', ('3408.26', '18.286201', 150, 2), ['18.560494', '63258.99', '1.50']),
  ('partial repair guard 1', ('419.33', '78.12165', 350, 2), ['80.855908', '33905.31', '3.51']),
  ('partial repair guard 2', ('2378.17', '122.42736', 300, 2), ['126.100181', '299887.67', '3.01']),
  ('control: euro to dollar', ('100.00', '1.083471', 300, 2), ['1.115975', '111.60', '3.00']),
  ('control: yen to dollar', ('5000', '0.00662', 350, 2), ['0.006852', '34.26', '3.51']),
  ('control: to yen card', ('49.99', '151.3742', 275, 0), ['155.536991', '7775', '2.76']),
  ('control: dinar card', ('250.00', '0.307133', 400, 3), ['0.319418', '79.855', '4.00'])]]
for label, args, expected in cases[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 fixture	Actual	Expected	Outcome
regression markup-disclosure-base 1	`['1.115975', '111.60', '3.00']`	`['1.115975', '111.60', '3.00']`	Passed
regression markup-disclosure-base 2	`['0.006852', '34.26', '3.50']`	`['0.006852', '34.26', '3.51']`	Failed
partial repair guard 2	`['155.536991', '7775', '2.75']`	`['155.536991', '7775', '2.76']`	Failed
control: dinar card	`['0.319418', '79.855', '4.00']`	`['0.319418', '79.855', '4.00']`	Passed
control: promotional discount	`['0.908810', '72.70', '-0.50']`	`['0.908810', '72.70', '-0.50']`	Passed
control: zero markup	`['1.200000', '14.81', '0.00']`	`['1.200000', '14.81', '0.00']`	Passed

SHA-256 / 02597a936818070cdab10313ac8617d0be26bceb34120b3026595d4a4b31837b

3 / The verified repair

Exit 0

"""Failure Map reference implementation. Python standard library only."""
import json
from decimal import Decimal, ROUND_HALF_EVEN, ROUND_HALF_UP, ROUND_FLOOR, ROUND_CEILING, ROUND_DOWN, ROUND_UP, ROUND_HALF_DOWN
N = 1
observations = []
def solve(amount, ref_rate, markup_bps, card_exp):
    ref = Decimal(ref_rate)
    raw = ref * (1 + Decimal(markup_bps) / 10000)
    dcc = raw.quantize(Decimal('0.000001'), rounding=ROUND_HALF_UP)
    card = (Decimal(amount) * dcc).quantize(Decimal(1).scaleb(-card_exp), rounding=ROUND_HALF_UP)
    pct = ((dcc / ref - 1) * 100).quantize(Decimal('0.01'), rounding=ROUND_CEILING)
    return [format(dcc, 'f'), format(card, 'f'), format(pct, 'f')]
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
cases = [[('regression markup-disclosure-base 1', ('100.00', '1.083471', 300, 2), ['1.115975', '111.60', '3.00']),
  ('regression markup-disclosure-base 2', ('5000', '0.00662', 350, 2), ['0.006852', '34.26', '3.51']),
  ('partial repair guard 2', ('49.99', '151.3742', 275, 0), ['155.536991', '7775', '2.76']),
  ('control: dinar card', ('250.00', '0.307133', 400, 3), ['0.319418', '79.855', '4.00']),
  ('control: promotional discount', ('80.00', '0.913377', -50, 2), ['0.908810', '72.70', '-0.50']),
  ('control: zero markup', ('12.34', '1.2', 0, 2), ['1.200000', '14.81', '0.00'])],
 [('regression markup-disclosure-base 1', ('49.99', '151.3742', 275, 0), ['155.536991', '7775', '2.76']),
  ('regression markup-disclosure-base 2', ('250.00', '0.307133', 400, 3), ['0.319418', '79.855', '4.00']),
  ('partial repair guard 1', ('3353.43', '37.36650', 275, 0), ['38.394079', '128752', '2.76']),
  ('partial repair guard 2', ('1171.26', '0.4683115', -50, 3), ['0.465970', '545.772', '-0.49']),
  ('control: promotional discount', ('80.00', '0.913377', -50, 2), ['0.908810', '72.70', '-0.50']),
  ('control: zero markup', ('12.34', '1.2', 0, 2), ['1.200000', '14.81', '0.00']),
  ('control: euro to dollar', ('100.00', '1.083471', 300, 2), ['1.115975', '111.60', '3.00']),
  ('control: yen to dollar', ('5000', '0.00662', 350, 2), ['0.006852', '34.26', '3.51'])],
 [('regression markup-disclosure-base 1', ('3353.43', '37.36650', 275, 0), ['38.394079', '128752', '2.76']),
  ('regression markup-disclosure-base 2', ('1171.26', '0.4683115', -50, 3), ['0.465970', '545.772', '-0.49']),
  ('partial repair guard 1', ('157.22', '31.44890', 150, 3), ['31.920634', '5018.562', '1.51']),
  ('partial repair guard 2', ('3157.31', '0.0891399', 150, 3), ['0.090477', '285.664', '1.51']),
  ('control: euro to dollar', ('100.00', '1.083471', 300, 2), ['1.115975', '111.60', '3.00']),
  ('control: yen to dollar', ('5000', '0.00662', 350, 2), ['0.006852', '34.26', '3.51']),
  ('control: to yen card', ('49.99', '151.3742', 275, 0), ['155.536991', '7775', '2.76']),
  ('control: dinar card', ('250.00', '0.307133', 400, 3), ['0.319418', '79.855', '4.00'])],
 [('regression markup-disclosure-base 1', ('3884.23', '183.00464', 350, 0), ['189.409802', '735711', '3.50']),
  ('regression markup-disclosure-base 2', ('157.22', '31.44890', 150, 3), ['31.920634', '5018.562', '1.51']),
  ('partial repair guard 1', ('508.07', '12.046349', 275, 0), ['12.377624', '6289', '2.76']),
  ('partial repair guard 2', ('3826.89', '1.5416740', -50, 2), ['1.533966', '5870.32', '-0.49']),
  ('control: dinar card', ('250.00', '0.307133', 400, 3), ['0.319418', '79.855', '4.00']),
  ('control: promotional discount', ('80.00', '0.913377', -50, 2), ['0.908810', '72.70', '-0.50']),
  ('control: zero markup', ('12.34', '1.2', 0, 2), ['1.200000', '14.81', '0.00']),
  ('control: euro to dollar', ('100.00', '1.083471', 300, 2), ['1.115975', '111.60', '3.00'])],
 [('regression markup-disclosure-base 1', ('3157.31', '0.0891399', 150, 3), ['0.090477', '285.664', '1.51']),
  ('regression markup-disclosure-base 2', ('3408.26', '18.286201', 150, 2), ['18.560494', '63258.99', '1.50']),
  ('partial repair guard 1', ('419.33', '78.12165', 350, 2), ['80.855908', '33905.31', '3.51']),
  ('partial repair guard 2', ('2378.17', '122.42736', 300, 2), ['126.100181', '299887.67', '3.01']),
  ('control: euro to dollar', ('100.00', '1.083471', 300, 2), ['1.115975', '111.60', '3.00']),
  ('control: yen to dollar', ('5000', '0.00662', 350, 2), ['0.006852', '34.26', '3.51']),
  ('control: to yen card', ('49.99', '151.3742', 275, 0), ['155.536991', '7775', '2.76']),
  ('control: dinar card', ('250.00', '0.307133', 400, 3), ['0.319418', '79.855', '4.00'])]]
for label, args, expected in cases[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 fixture	Actual	Expected	Outcome
regression markup-disclosure-base 1	`['1.115975', '111.60', '3.00']`	`['1.115975', '111.60', '3.00']`	Passed
regression markup-disclosure-base 2	`['0.006852', '34.26', '3.51']`	`['0.006852', '34.26', '3.51']`	Passed
partial repair guard 2	`['155.536991', '7775', '2.76']`	`['155.536991', '7775', '2.76']`	Passed
control: dinar card	`['0.319418', '79.855', '4.00']`	`['0.319418', '79.855', '4.00']`	Passed
control: promotional discount	`['0.908810', '72.70', '-0.50']`	`['0.908810', '72.70', '-0.50']`	Passed
control: zero markup	`['1.200000', '14.81', '0.00']`	`['1.200000', '14.81', '0.00']`	Passed

SHA-256 / e3eb2b9a5280501fc53d61bbacc448f34101395525f37605e0e74e02a4f4c6a2

Verification & scope

A deterministic, bounded teaching model with a stipulated toy contract; it makes no claim of conformance to any real regulation, standard, or institution's rules. 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 2026-09-29T14:47:02.317626+00:00.

Case digest / cd7c618ecb836551104d3b6847060e3eb54f439ec557bfd1c411c0ff36d756d9