FAILURE MAP
← Case archive

FA-86466 / Procedural level generation constraints / Open access

Platform jump reachability: Drops extend horizontal reach · case 01

Falling platforms become reachable from absurd distances.

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

ROOT CAUSE

Negative rises reduce the penalty, adding reach on drops.

VERIFIED REPAIR

Restore `jump_d - max(0, rise) // 2` at the climb penalty step.

Unsuccessful approach: Penalising drops as well as climbs removes legal drop routes.

Case contract

platforms [x1, x2, y] (y grows upward); the player starts on platform 0. From A the player reaches B when rise = yB-yA <= jump_h (any drop is allowed) and the horizontal gap max(0, xB1-xA2, xA1-xB2) <= jump_d - max(0, rise)//2. Reachability is transitive. Returns sorted reachable indices.

Why this case matters

Procedural generators silently emit unplayable or unfair levels when a single constraint check uses the wrong boundary, axis, neighborhood or update order; the defect is visible in exact generated geometry.

1 / The failure

Exit 1
"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(platforms, jump_h, jump_d):
    n = len(platforms)
    seen = {0}
    queue = [0]
    while queue:
        a = queue.pop(0)
        ax1, ax2, ay = platforms[a]
        for b in range(n):
            if b in seen:
                continue
            bx1, bx2, by = platforms[b]
            rise = by - ay
            if rise > jump_h:
                continue
            gap = max(0, bx1 - ax2, ax1 - bx2)
            reach = jump_d - rise // 2
            if gap <= reach:
                seen.add(b)
                queue.append(b)
    return sorted(seen)
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
cases = [[('deep drop #1', [[[0, 2, 10], [3, 5, 0]], 2, 2], [0, 1]),
  ('regression climb penalty #1',
   [[[23, 25, 6], [15, 16, 2], [21, 21, 7], [27, 33, 3], [2, 8, 3]], 2, 2],
   [0, 2, 3]),
  ('fault site climb penalty #1', [[[20, 24, 10], [6, 8, 1], [30, 30, 5]], 5, 5], [0]),
  ('partial repair boundary #1',
   [[[5, 9, 8], [7, 13, 0], [0, 2, 9], [13, 13, 2], [25, 31, 3]], 2, 1],
   [0, 1, 3]),
  ('partial repair boundary #2', [[[23, 29, 6], [19, 22, 0], [30, 32, 8], [15, 20, 1]], 3, 1], [0, 1, 3]),
  ('platform to the left #1', [[[10, 12, 0], [4, 7, 0]], 1, 3], [0, 1]),
  ('chain of hops #1', [[[0, 1, 0], [3, 4, 2], [6, 7, 4]], 2, 3], [0, 1, 2]),
  ('control #1', [[[13, 18, 6], [20, 22, 7]], 2, 8], [0, 1])],
 [('deep drop #1', [[[0, 2, 10], [3, 5, 0]], 2, 2], [0, 1]),
  ('fault site climb penalty #1', [[[9, 14, 10], [0, 6, 1], [21, 21, 2]], 5, 2], [0]),
  ('fault site climb penalty #2',
   [[[0, 6, 6], [20, 25, 2], [5, 9, 8], [28, 31, 3], [7, 11, 8]], 3, 8],
   [0, 2, 4]),
  ('partial repair boundary #1', [[[23, 29, 6], [19, 22, 0], [30, 32, 8], [15, 20, 1]], 3, 1], [0, 1, 3]),
  ('regression climb penalty #1',
   [[[23, 25, 6], [15, 16, 2], [21, 21, 7], [27, 33, 3], [2, 8, 3]], 2, 2],
   [0, 2, 3]),
  ('platform to the left #1', [[[10, 12, 0], [4, 7, 0]], 1, 3], [0, 1]),
  ('chain of hops #1', [[[0, 1, 0], [3, 4, 2], [6, 7, 4]], 2, 3], [0, 1, 2]),
  ('control #1', [[[13, 18, 6], [20, 22, 7]], 2, 8], [0, 1])],
 [('deep drop #1', [[[0, 2, 10], [3, 5, 0]], 2, 2], [0, 1]),
  ('fault site climb penalty #1',
   [[[24, 26, 1], [7, 10, 9], [5, 7, 8], [21, 27, 5], [18, 19, 0], [29, 32, 9]], 1, 4],
   [0]),
  ('fault site climb penalty #2',
   [[[17, 17, 4], [11, 11, 6], [26, 32, 4], [14, 19, 5], [28, 33, 4], [27, 29, 2]], 5, 6],
   [0, 1, 3]),
  ('partial repair boundary #1',
   [[[18, 18, 5], [15, 19, 6], [21, 22, 3], [5, 9, 1], [14, 20, 6], [24, 29, 7]], 1, 5],
   [0, 1, 2, 3, 4, 5]),
  ('partial repair boundary #2',
   [[[7, 9, 9], [10, 15, 0], [7, 10, 2], [9, 14, 4], [26, 30, 8], [15, 19, 1]], 5, 1],
   [0, 1, 2, 3, 5]),
  ('platform to the left #1', [[[10, 12, 0], [4, 7, 0]], 1, 3], [0, 1]),
  ('chain of hops #1', [[[0, 1, 0], [3, 4, 2], [6, 7, 4]], 2, 3], [0, 1, 2]),
  ('control #1', [[[13, 18, 6], [20, 22, 7]], 2, 8], [0, 1])],
 [('deep drop #1', [[[0, 2, 10], [3, 5, 0]], 2, 2], [0, 1]),
  ('fault site climb penalty #1',
   [[[4, 9, 10], [0, 5, 1], [27, 30, 1], [18, 21, 5], [19, 25, 6]], 4, 8],
   [0, 1]),
  ('fault site climb penalty #2', [[[10, 10, 10], [21, 22, 10], [16, 22, 2], [3, 3, 5]], 1, 2], [0]),
  ('partial repair boundary #1',
   [[[5, 10, 6], [19, 21, 1], [6, 7, 0], [11, 17, 10], [29, 35, 5], [0, 0, 8]], 5, 1],
   [0, 2]),
  ('partial repair boundary #2',
   [[[8, 9, 9], [9, 14, 3], [9, 9, 2], [5, 8, 10], [10, 14, 4], [24, 26, 6]], 5, 2],
   [0, 1, 2, 3, 4]),
  ('platform to the left #1', [[[10, 12, 0], [4, 7, 0]], 1, 3], [0, 1]),
  ('chain of hops #1', [[[0, 1, 0], [3, 4, 2], [6, 7, 4]], 2, 3], [0, 1, 2]),
  ('control #1', [[[23, 29, 1], [3, 9, 0]], 4, 3], [0])],
 [('deep drop #1', [[[0, 2, 10], [3, 5, 0]], 2, 2], [0, 1]),
  ('fault site climb penalty #1', [[[20, 23, 7], [24, 26, 9], [25, 25, 6], [7, 12, 0]], 5, 5], [0, 1, 2]),
  ('fault site climb penalty #2',
   [[[12, 16, 8], [30, 34, 6], [15, 15, 8], [24, 25, 1], [25, 29, 6], [30, 35, 7]], 2, 7],
   [0, 2]),
  ('partial repair boundary #1',
   [[[20, 26, 10], [29, 29, 4], [18, 18, 3], [17, 18, 0], [16, 22, 8]], 5, 4],
   [0, 1, 2, 3, 4]),
  ('partial repair boundary #2',
   [[[20, 25, 6], [15, 17, 9], [23, 24, 0], [21, 23, 4], [11, 13, 8], [13, 17, 0]], 3, 3],
   [0, 2, 3, 5]),
  ('platform to the left #1', [[[10, 12, 0], [4, 7, 0]], 1, 3], [0, 1]),
  ('chain of hops #1', [[[0, 1, 0], [3, 4, 2], [6, 7, 4]], 2, 3], [0, 1, 2]),
  ('control #1', [[[17, 20, 8]], 4, 4], [0])]]
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 fixtureActualExpectedOutcome
deep drop #1[0, 1][0, 1]Passed
regression climb penalty #1[0, 1, 2, 3][0, 2, 3]Failed
fault site climb penalty #1[0, 2][0]Failed
partial repair boundary #1[0, 1, 3][0, 1, 3]Passed
partial repair boundary #2[0, 1, 3][0, 1, 3]Passed
platform to the left #1[0, 1][0, 1]Passed
chain of hops #1[0, 1, 2][0, 1, 2]Passed
control #1[0, 1][0, 1]Passed

SHA-256 / 7a1c019c51112eba0696c5a59ee9a5b6f4b6065b3b53c41a3dc897e651110910

2 / The unsuccessful fix

Exit 1
"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(platforms, jump_h, jump_d):
    n = len(platforms)
    seen = {0}
    queue = [0]
    while queue:
        a = queue.pop(0)
        ax1, ax2, ay = platforms[a]
        for b in range(n):
            if b in seen:
                continue
            bx1, bx2, by = platforms[b]
            rise = by - ay
            if rise > jump_h:
                continue
            gap = max(0, bx1 - ax2, ax1 - bx2)
            reach = jump_d - abs(rise) // 2
            if gap <= reach:
                seen.add(b)
                queue.append(b)
    return sorted(seen)
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
cases = [[('deep drop #1', [[[0, 2, 10], [3, 5, 0]], 2, 2], [0, 1]),
  ('regression climb penalty #1',
   [[[23, 25, 6], [15, 16, 2], [21, 21, 7], [27, 33, 3], [2, 8, 3]], 2, 2],
   [0, 2, 3]),
  ('fault site climb penalty #1', [[[20, 24, 10], [6, 8, 1], [30, 30, 5]], 5, 5], [0]),
  ('partial repair boundary #1',
   [[[5, 9, 8], [7, 13, 0], [0, 2, 9], [13, 13, 2], [25, 31, 3]], 2, 1],
   [0, 1, 3]),
  ('partial repair boundary #2', [[[23, 29, 6], [19, 22, 0], [30, 32, 8], [15, 20, 1]], 3, 1], [0, 1, 3]),
  ('platform to the left #1', [[[10, 12, 0], [4, 7, 0]], 1, 3], [0, 1]),
  ('chain of hops #1', [[[0, 1, 0], [3, 4, 2], [6, 7, 4]], 2, 3], [0, 1, 2]),
  ('control #1', [[[13, 18, 6], [20, 22, 7]], 2, 8], [0, 1])],
 [('deep drop #1', [[[0, 2, 10], [3, 5, 0]], 2, 2], [0, 1]),
  ('fault site climb penalty #1', [[[9, 14, 10], [0, 6, 1], [21, 21, 2]], 5, 2], [0]),
  ('fault site climb penalty #2',
   [[[0, 6, 6], [20, 25, 2], [5, 9, 8], [28, 31, 3], [7, 11, 8]], 3, 8],
   [0, 2, 4]),
  ('partial repair boundary #1', [[[23, 29, 6], [19, 22, 0], [30, 32, 8], [15, 20, 1]], 3, 1], [0, 1, 3]),
  ('regression climb penalty #1',
   [[[23, 25, 6], [15, 16, 2], [21, 21, 7], [27, 33, 3], [2, 8, 3]], 2, 2],
   [0, 2, 3]),
  ('platform to the left #1', [[[10, 12, 0], [4, 7, 0]], 1, 3], [0, 1]),
  ('chain of hops #1', [[[0, 1, 0], [3, 4, 2], [6, 7, 4]], 2, 3], [0, 1, 2]),
  ('control #1', [[[13, 18, 6], [20, 22, 7]], 2, 8], [0, 1])],
 [('deep drop #1', [[[0, 2, 10], [3, 5, 0]], 2, 2], [0, 1]),
  ('fault site climb penalty #1',
   [[[24, 26, 1], [7, 10, 9], [5, 7, 8], [21, 27, 5], [18, 19, 0], [29, 32, 9]], 1, 4],
   [0]),
  ('fault site climb penalty #2',
   [[[17, 17, 4], [11, 11, 6], [26, 32, 4], [14, 19, 5], [28, 33, 4], [27, 29, 2]], 5, 6],
   [0, 1, 3]),
  ('partial repair boundary #1',
   [[[18, 18, 5], [15, 19, 6], [21, 22, 3], [5, 9, 1], [14, 20, 6], [24, 29, 7]], 1, 5],
   [0, 1, 2, 3, 4, 5]),
  ('partial repair boundary #2',
   [[[7, 9, 9], [10, 15, 0], [7, 10, 2], [9, 14, 4], [26, 30, 8], [15, 19, 1]], 5, 1],
   [0, 1, 2, 3, 5]),
  ('platform to the left #1', [[[10, 12, 0], [4, 7, 0]], 1, 3], [0, 1]),
  ('chain of hops #1', [[[0, 1, 0], [3, 4, 2], [6, 7, 4]], 2, 3], [0, 1, 2]),
  ('control #1', [[[13, 18, 6], [20, 22, 7]], 2, 8], [0, 1])],
 [('deep drop #1', [[[0, 2, 10], [3, 5, 0]], 2, 2], [0, 1]),
  ('fault site climb penalty #1',
   [[[4, 9, 10], [0, 5, 1], [27, 30, 1], [18, 21, 5], [19, 25, 6]], 4, 8],
   [0, 1]),
  ('fault site climb penalty #2', [[[10, 10, 10], [21, 22, 10], [16, 22, 2], [3, 3, 5]], 1, 2], [0]),
  ('partial repair boundary #1',
   [[[5, 10, 6], [19, 21, 1], [6, 7, 0], [11, 17, 10], [29, 35, 5], [0, 0, 8]], 5, 1],
   [0, 2]),
  ('partial repair boundary #2',
   [[[8, 9, 9], [9, 14, 3], [9, 9, 2], [5, 8, 10], [10, 14, 4], [24, 26, 6]], 5, 2],
   [0, 1, 2, 3, 4]),
  ('platform to the left #1', [[[10, 12, 0], [4, 7, 0]], 1, 3], [0, 1]),
  ('chain of hops #1', [[[0, 1, 0], [3, 4, 2], [6, 7, 4]], 2, 3], [0, 1, 2]),
  ('control #1', [[[23, 29, 1], [3, 9, 0]], 4, 3], [0])],
 [('deep drop #1', [[[0, 2, 10], [3, 5, 0]], 2, 2], [0, 1]),
  ('fault site climb penalty #1', [[[20, 23, 7], [24, 26, 9], [25, 25, 6], [7, 12, 0]], 5, 5], [0, 1, 2]),
  ('fault site climb penalty #2',
   [[[12, 16, 8], [30, 34, 6], [15, 15, 8], [24, 25, 1], [25, 29, 6], [30, 35, 7]], 2, 7],
   [0, 2]),
  ('partial repair boundary #1',
   [[[20, 26, 10], [29, 29, 4], [18, 18, 3], [17, 18, 0], [16, 22, 8]], 5, 4],
   [0, 1, 2, 3, 4]),
  ('partial repair boundary #2',
   [[[20, 25, 6], [15, 17, 9], [23, 24, 0], [21, 23, 4], [11, 13, 8], [13, 17, 0]], 3, 3],
   [0, 2, 3, 5]),
  ('platform to the left #1', [[[10, 12, 0], [4, 7, 0]], 1, 3], [0, 1]),
  ('chain of hops #1', [[[0, 1, 0], [3, 4, 2], [6, 7, 4]], 2, 3], [0, 1, 2]),
  ('control #1', [[[17, 20, 8]], 4, 4], [0])]]
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 fixtureActualExpectedOutcome
deep drop #1[0][0, 1]Failed
regression climb penalty #1[0, 2][0, 2, 3]Failed
fault site climb penalty #1[0][0]Passed
partial repair boundary #1[0][0, 1, 3]Failed
partial repair boundary #2[0][0, 1, 3]Failed
platform to the left #1[0, 1][0, 1]Passed
chain of hops #1[0, 1, 2][0, 1, 2]Passed
control #1[0, 1][0, 1]Passed

SHA-256 / 9cd0f77a2e2874f4d13bbf496cd5258f1d126cfd8bb61c697b9edb417b5ab1e9

3 / The verified repair

Exit 0
"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(platforms, jump_h, jump_d):
    n = len(platforms)
    seen = {0}
    queue = [0]
    while queue:
        a = queue.pop(0)
        ax1, ax2, ay = platforms[a]
        for b in range(n):
            if b in seen:
                continue
            bx1, bx2, by = platforms[b]
            rise = by - ay
            if rise > jump_h:
                continue
            gap = max(0, bx1 - ax2, ax1 - bx2)
            reach = jump_d - max(0, rise) // 2
            if gap <= reach:
                seen.add(b)
                queue.append(b)
    return sorted(seen)
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
cases = [[('deep drop #1', [[[0, 2, 10], [3, 5, 0]], 2, 2], [0, 1]),
  ('regression climb penalty #1',
   [[[23, 25, 6], [15, 16, 2], [21, 21, 7], [27, 33, 3], [2, 8, 3]], 2, 2],
   [0, 2, 3]),
  ('fault site climb penalty #1', [[[20, 24, 10], [6, 8, 1], [30, 30, 5]], 5, 5], [0]),
  ('partial repair boundary #1',
   [[[5, 9, 8], [7, 13, 0], [0, 2, 9], [13, 13, 2], [25, 31, 3]], 2, 1],
   [0, 1, 3]),
  ('partial repair boundary #2', [[[23, 29, 6], [19, 22, 0], [30, 32, 8], [15, 20, 1]], 3, 1], [0, 1, 3]),
  ('platform to the left #1', [[[10, 12, 0], [4, 7, 0]], 1, 3], [0, 1]),
  ('chain of hops #1', [[[0, 1, 0], [3, 4, 2], [6, 7, 4]], 2, 3], [0, 1, 2]),
  ('control #1', [[[13, 18, 6], [20, 22, 7]], 2, 8], [0, 1])],
 [('deep drop #1', [[[0, 2, 10], [3, 5, 0]], 2, 2], [0, 1]),
  ('fault site climb penalty #1', [[[9, 14, 10], [0, 6, 1], [21, 21, 2]], 5, 2], [0]),
  ('fault site climb penalty #2',
   [[[0, 6, 6], [20, 25, 2], [5, 9, 8], [28, 31, 3], [7, 11, 8]], 3, 8],
   [0, 2, 4]),
  ('partial repair boundary #1', [[[23, 29, 6], [19, 22, 0], [30, 32, 8], [15, 20, 1]], 3, 1], [0, 1, 3]),
  ('regression climb penalty #1',
   [[[23, 25, 6], [15, 16, 2], [21, 21, 7], [27, 33, 3], [2, 8, 3]], 2, 2],
   [0, 2, 3]),
  ('platform to the left #1', [[[10, 12, 0], [4, 7, 0]], 1, 3], [0, 1]),
  ('chain of hops #1', [[[0, 1, 0], [3, 4, 2], [6, 7, 4]], 2, 3], [0, 1, 2]),
  ('control #1', [[[13, 18, 6], [20, 22, 7]], 2, 8], [0, 1])],
 [('deep drop #1', [[[0, 2, 10], [3, 5, 0]], 2, 2], [0, 1]),
  ('fault site climb penalty #1',
   [[[24, 26, 1], [7, 10, 9], [5, 7, 8], [21, 27, 5], [18, 19, 0], [29, 32, 9]], 1, 4],
   [0]),
  ('fault site climb penalty #2',
   [[[17, 17, 4], [11, 11, 6], [26, 32, 4], [14, 19, 5], [28, 33, 4], [27, 29, 2]], 5, 6],
   [0, 1, 3]),
  ('partial repair boundary #1',
   [[[18, 18, 5], [15, 19, 6], [21, 22, 3], [5, 9, 1], [14, 20, 6], [24, 29, 7]], 1, 5],
   [0, 1, 2, 3, 4, 5]),
  ('partial repair boundary #2',
   [[[7, 9, 9], [10, 15, 0], [7, 10, 2], [9, 14, 4], [26, 30, 8], [15, 19, 1]], 5, 1],
   [0, 1, 2, 3, 5]),
  ('platform to the left #1', [[[10, 12, 0], [4, 7, 0]], 1, 3], [0, 1]),
  ('chain of hops #1', [[[0, 1, 0], [3, 4, 2], [6, 7, 4]], 2, 3], [0, 1, 2]),
  ('control #1', [[[13, 18, 6], [20, 22, 7]], 2, 8], [0, 1])],
 [('deep drop #1', [[[0, 2, 10], [3, 5, 0]], 2, 2], [0, 1]),
  ('fault site climb penalty #1',
   [[[4, 9, 10], [0, 5, 1], [27, 30, 1], [18, 21, 5], [19, 25, 6]], 4, 8],
   [0, 1]),
  ('fault site climb penalty #2', [[[10, 10, 10], [21, 22, 10], [16, 22, 2], [3, 3, 5]], 1, 2], [0]),
  ('partial repair boundary #1',
   [[[5, 10, 6], [19, 21, 1], [6, 7, 0], [11, 17, 10], [29, 35, 5], [0, 0, 8]], 5, 1],
   [0, 2]),
  ('partial repair boundary #2',
   [[[8, 9, 9], [9, 14, 3], [9, 9, 2], [5, 8, 10], [10, 14, 4], [24, 26, 6]], 5, 2],
   [0, 1, 2, 3, 4]),
  ('platform to the left #1', [[[10, 12, 0], [4, 7, 0]], 1, 3], [0, 1]),
  ('chain of hops #1', [[[0, 1, 0], [3, 4, 2], [6, 7, 4]], 2, 3], [0, 1, 2]),
  ('control #1', [[[23, 29, 1], [3, 9, 0]], 4, 3], [0])],
 [('deep drop #1', [[[0, 2, 10], [3, 5, 0]], 2, 2], [0, 1]),
  ('fault site climb penalty #1', [[[20, 23, 7], [24, 26, 9], [25, 25, 6], [7, 12, 0]], 5, 5], [0, 1, 2]),
  ('fault site climb penalty #2',
   [[[12, 16, 8], [30, 34, 6], [15, 15, 8], [24, 25, 1], [25, 29, 6], [30, 35, 7]], 2, 7],
   [0, 2]),
  ('partial repair boundary #1',
   [[[20, 26, 10], [29, 29, 4], [18, 18, 3], [17, 18, 0], [16, 22, 8]], 5, 4],
   [0, 1, 2, 3, 4]),
  ('partial repair boundary #2',
   [[[20, 25, 6], [15, 17, 9], [23, 24, 0], [21, 23, 4], [11, 13, 8], [13, 17, 0]], 3, 3],
   [0, 2, 3, 5]),
  ('platform to the left #1', [[[10, 12, 0], [4, 7, 0]], 1, 3], [0, 1]),
  ('chain of hops #1', [[[0, 1, 0], [3, 4, 2], [6, 7, 4]], 2, 3], [0, 1, 2]),
  ('control #1', [[[17, 20, 8]], 4, 4], [0])]]
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 fixtureActualExpectedOutcome
deep drop #1[0, 1][0, 1]Passed
regression climb penalty #1[0, 2, 3][0, 2, 3]Passed
fault site climb penalty #1[0][0]Passed
partial repair boundary #1[0, 1, 3][0, 1, 3]Passed
partial repair boundary #2[0, 1, 3][0, 1, 3]Passed
platform to the left #1[0, 1][0, 1]Passed
chain of hops #1[0, 1, 2][0, 1, 2]Passed
control #1[0, 1][0, 1]Passed

SHA-256 / 3146d701158af3203807f815f98b6c0278e8f13c87808e29b1254b352bb2e099

Verification & scope

Deterministic toy contract stipulated for this model; integer or exact arithmetic only, not a reproduction of any specific game engine. 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 2026-09-29T14:50:49.917221+00:00.

Case digest / adbdf914afb3d5e35a4636f9df88e2740de38d69d64c13880e50ba3a3c1d852b