关于「格子宽度与多层盘面:刻、时辰与上升星座」,《出生时刻误差与交界翻转》给出的定量结论是什么?
结论:第三章把格子扩展到刻、整点与上升星座,几层叠加后有 0.7230 的人至少有一层盘面不符。
数据口径:本书只分析排盘规则的结构、误差传播与概率,不做任何个人占断,不宣称任何体系有预测效力;书中参数为示意,全部计算由书中程序生成。 第一章说明,示意模型中凭记忆报告的时刻有 0.2782 恰好落在时辰分界上,不取整时只有 0.0083。
对照数据表
下表数字由本章 Python 程序原样打印,可重新运行复现。 程序只用 Python 标准库与结论吻合的打印值: 0.7230
| share of people whose reported chart differs from the true one | |
| double hour | 0.1736 |
| clock hour | 0.3471 |
| quarter hour | 0.7076 |
| rising sign | 0.1259 |
| any of the four | 0.7230 |
| double hour or rising sign, if independent: 0.2776 | |
| day-pillar convention split (23:00 vs 00:00): 0.0433 |
核心公式
\[P_{\text{flip}} = 1 - \frac{\sigma}{w}\left[x\,(2\Phi(x)-1) + 2\varphi(x) - 2\varphi(0)\right],\qquad x = w/\sigma .\]
查看完整验算程序
# Chapter 3. p33: a chart has several layers, and any one can flip.
# Fixed-seed Monte Carlo of the memory source of p11 (illustrative
# weights, recall error 15 minutes). For each simulated person we
# compare the cell of the true time and of the report in four layers:
# double hour (boundaries at odd hours), clock hour, quarter hour, and
# the rising sign at latitude 31.2 (sign lengths from p32; the sky's
# phase at midnight is drawn at random, so its boundaries are not tied
# to the clock). We also count a pure convention split: the day pillar
# changes at 23:00 in one school and at 00:00 in another, so births in
# 23:00-24:00 get different day pillars with no error at all.
import bisect
import math
import random
GRAN = [1, 5, 15, 30, 60]
W, S = [0.10, 0.15, 0.10, 0.25, 0.40], 15.0
# 纬度 31.2 度的十二宫上升时长(时钟分钟,取自 p32)
SIGN_MIN = [83.2, 96.2, 118.9, 137.9, 142.4, 139.5, 139.5, 142.4, 137.9,
118.9, 96.2, 83.2]
DAY = math.fsum(SIGN_MIN)
EDGES = [math.fsum(SIGN_MIN[:k + 1]) for k in range(12)]
def sign_at(t, phase):
x = (t * DAY / 1440.0 + phase) % DAY
return bisect.bisect_right(EDGES, x)
rng = random.Random(20261113)
N = 100000
hits = {"double hour": 0, "clock hour": 0, "quarter hour": 0,
"rising sign": 0, "any of the four": 0}
conv = 0
for _ in range(N):
t = rng.random() * 1440.0
g = rng.choices(GRAN, W)[0]
r = (math.floor((t + rng.gauss(0.0, S)) / g + 0.5) * g) % 1440
phase = rng.random() * DAY
f = [int((t + 60) % 1440 // 120) != int((r + 60) % 1440 // 120),
int(t // 60) != int(r // 60),
int(t // 15) != int(r // 15),
sign_at(t, phase) != sign_at(r, phase)]
for name, x in zip(("double hour", "clock hour", "quarter hour",
"rising sign"), f):
hits[name] += x # step
hits["any of the four"] += any(f) # step
conv += t >= 1380.0 # step
print("share of people whose reported chart differs from the true one")
for name in ("double hour", "clock hour", "quarter hour", "rising sign",
"any of the four"):
print(" %-17s %.4f" % (name, hits[name] / N))
print("double hour or rising sign, if independent: %.4f"
% (1.0 - (1.0 - hits["double hour"] / N)
* (1.0 - hits["rising sign"] / N)))
print("day-pillar convention split (23:00 vs 00:00): %.4f" % (conv / N))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《出生时刻记忆误差传播与鲁棒定盘、交界时辰盘面翻转概率与最少前事反推题数母典》第 3 章「格子宽度与多层盘面:刻、时辰与上升星座」(版本 1.0.0). 5D Library. DOI 待发布
DOI 待发布 · 专著页正式发布后上线(本站暂无专著页) · 5DL-B-284 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026284,
author = {Fan, Yafu},
title = {Error Propagation from Remembered Birth Times and Robust Chart Fixing: Flip Probabilities at Double-Hour Boundaries and the Fewest Past-Event Questions},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}