5D Library › 姓名笔画、生命灵数与手机号码的同余映射和分布论 · English
第二章 生命灵数在人群中的基率
为什么要先算基率
一条断语若对九分之一的人成立,与一条对三分之一的人成立的断语,「说中」的意义完全不同。要评价任何一种分类,第一步是知道每一类在人群中占多少,这就是基率 (Kahneman and Tversky 1973)。生命灵数的基率常被默认为九分之一,因为余数有九个。本章检查这一默认。
公历的月份有 28 到 31 天,闰年多一天,所以出生日期并不是均匀地分布在余数上的一个集合。下面的程序假定每个日历日的出生人数相同,计算三个出生人群中生命灵数的比例:1950 年到 2009 年、1990 年代、单独的 1995 年。真实的出生有季节与星期的起伏,这只会在下面的不均匀之上再加一层。
推导
# Chapter 2. Base rates of life path numbers among people.
# Births are taken as equally likely on every calendar day of the stated
# years (an assumption: real births have a seasonal pattern and weekday
# effects, which only add to the unevenness found here). Convention A of
# the previous program (all digits, stop at 11/22/33) and the plain
# digital root D are compared for three populations.
import datetime as dt
MASTERS = (11, 22, 33)
LABELS = list(range(1, 10)) + [11, 22, 33]
def digit_sum(n):
return sum(int(c) for c in str(n))
def life_path(y, m, d):
n = digit_sum(y) + digit_sum(m) + digit_sum(d)
while n > 9 and n not in MASTERS:
n = digit_sum(n)
return n
def shares(y0, y1, rule):
count, n = {k: 0 for k in LABELS}, 0
day, end = dt.date(y0, 1, 1), dt.date(y1, 12, 31)
while day <= end:
count[rule(day.year, day.month, day.day)] += 1
n += 1
day += dt.timedelta(days=1)
return {k: c / n for k, c in count.items()}
def plain(y, m, d):
return 1 + (y + m + d - 1) % 9
if __name__ == "__main__":
print(" share of people, per cent (births equally likely each day)")
for y0, y1 in ((1950, 2009), (1990, 1999), (1995, 1995)):
a, d = shares(y0, y1, life_path), shares(y0, y1, plain)
print(" births {}-{}".format(y0, y1))
print(" number " + "".join("{:>5}".format(k) for k in LABELS))
print(" rule A " + "".join("{:5.1f}".format(100 * a[k])
for k in LABELS))
print(" plain " + "".join("{:5.1f}".format(100 * d[k])
for k in LABELS[:9]))
core = [a[k] for k in range(1, 10)]
rare = min(range(1, 10), key=lambda k: a[k])
print(" rule A: rarest of 1-9 is {} at {:.2f} per cent;"
" largest/rarest = {}".format(
rare, 100 * a[rare],
"{:.1f}".format(max(core) / a[rare]) if a[rare] else
"infinite"))
print(" uniform share 1/9 = {:.1f} per cent".format(100 / 9))
print("\nReading: over sixty or ten birth years the plain residues are")
print("even to one decimal; within one year they spread from 10.7 to")
print("11.5. The master rule takes its 11s, 22s and 33s from 2, 4 and 6,")
print("so those become the rarest; for births in the 1990s no date at")
print("all gives a life path of 2.")
# share of people, per cent (births equally likely each day)
# births 1950-2009
# number 1 2 3 4 5 6 7 8 9 11 22 33
# rule A 11.1 1.5 11.1 9.1 11.1 5.2 11.1 11.1 11.1 9.6 2.0 5.9
# plain 11.1 11.1 11.1 11.1 11.1 11.1 11.1 11.1 11.1
# rule A: rarest of 1-9 is 2 at 1.49 per cent; largest/rarest = 7.5
# births 1990-1999
# number 1 2 3 4 5 6 7 8 9 11 22 33
# rule A 11.1 0.0 11.1 10.8 11.1 3.3 11.1 11.1 11.1 11.1 0.4 7.9
# plain 11.1 11.1 11.1 11.1 11.1 11.1 11.1 11.1 11.1
# rule A: rarest of 1-9 is 2 at 0.00 per cent; largest/rarest = infinite
# births 1995-1995
# number 1 2 3 4 5 6 7 8 9 11 22 33
# rule A 11.2 0.0 11.0 11.2 10.7 1.6 10.7 11.2 11.5 11.5 0.0 9.3
# plain 11.2 11.5 11.0 11.2 10.7 11.0 10.7 11.2 11.5
# rule A: rarest of 1-9 is 2 at 0.00 per cent; largest/rarest = infinite
# uniform share 1/9 = 11.1 per cent
#
# Reading: over sixty or ten birth years the plain residues are
# even to one decimal; within one year they spread from 10.7 to
# 11.5. The master rule takes its 11s, 22s and 33s from 2, 4 and 6,
# so those become the rarest; for births in the 1990s no date at
# all gives a life path of 2.余数几乎均匀,大师数不均匀
结果分成两部分。纯数字根在六十年与十年的人群中,每个余数都是 11.1,即九分之一,到小数点后一位都一样;只看一年出生的人,余数的比例在 10.7 到 11.5 之间。月份长短不一,并没有在余数上留下可观的偏倚。
大师数则完全不同。按第一章的算法 A,11 从余数 2 那里取走人,22 从 4 那里取走,33 从 6 那里取走。在 1950 年到 2009 年出生的人中,生命灵数 2 只占 1.5,22 只占 2.0,而 11 占 9.6;1 到 9 中最常见与最少见之比为 7.5。到了 1990 年代,没有一个日期给出生命灵数 2:这些年份的数字 1、9、9 已经占去 19,再加上月与日的数字,和至少是 21,余 2 的和只能是 29、38 或 47,它们全都停在 11。
注
同一个「2 号人」在不同年代的意义。 按算法 A,「生命灵数 2」在 1950 年代出生的人中尚有一些,在 1990 年代出生的人中则一个也没有。任何关于「2 号人」的人群统计,都首先是关于出生年代与算法的统计。
流年数
流年数把出生的月与日加上当年的年份,再化简。年份只通过它的余数进入,所以对一个人来说,流年数每年加一,九年一轮;对整个人群来说,某一年里处于各个流年的人数比例,就是「月加日」余数的比例换了标签。
推导
# Chapter 2. The personal year number of a whole population.
# Personal year for calendar year Y = reduce(birth month + birth day + Y),
# by digits. Births are taken as equally likely on each day of the year,
# with 29 February weighted 1/4. Because Y enters only through its
# residue, the population's shares in a given year are the shares of
# (month + day) mod 9, rotated: the same nine numbers every year, only
# relabelled.
from fractions import Fraction as F
MASTERS = (11, 22)
DAYS = (31, 29, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31)
def digit_sum(n):
return sum(int(c) for c in str(n))
def red(n, masters=()):
while n > 9 and n not in masters:
n = digit_sum(n)
return n
def birthdays():
for m, nd in enumerate(DAYS, start=1):
for d in range(1, nd + 1):
yield m, d, (F(1, 4) if (m, d) == (2, 29) else F(1))
def year_shares(year, masters=()):
out, tot = {}, F(0)
for m, d, w in birthdays():
k = red(digit_sum(m) + digit_sum(d) + digit_sum(year), masters)
out[k] = out.get(k, 0) + w
tot += w
return {k: v / tot for k, v in sorted(out.items())}
if __name__ == "__main__":
keys = list(range(1, 10)) + [11, 22]
print(" share of people in each personal year, per cent")
print(" year rule " + "".join("{:>5}".format(k) for k in keys))
for year in (2026, 2027):
for label, ms in (("plain", ()), ("11/22", MASTERS)):
sh = year_shares(year, ms)
print(" {} {:5s}".format(year, label) + "".join(
"{:5.1f}".format(100 * float(sh.get(k, 0))) for k in keys))
p = year_shares(2026)
lo, hi = min(p.values()), max(p.values())
print(" plain shares: lowest {:.2f}, highest {:.2f}, ratio {:.3f}"
.format(100 * float(lo), 100 * float(hi), float(hi / lo)))
rot = all(year_shares(2026)[k] == year_shares(2027)[1 + k % 9]
for k in range(1, 10))
print(" 2027 shares are the 2026 shares moved up by one: {}"
.format(rot))
print("\nReading: without master numbers the nine personal years hold")
print("nearly equal shares of people, the same shares every year, only")
print("relabelled. With 11 and 22 kept, 22 takes most of the people")
print("who would be in year 4, while 11 takes only a few from year 2.")
# share of people in each personal year, per cent
# year rule 1 2 3 4 5 6 7 8 9 11 22
# 2026 plain 11.0 10.7 11.2 11.5 11.3 11.5 11.0 11.2 10.7 0.0 0.0
# 2026 11/22 11.0 9.9 11.2 2.7 11.3 11.5 11.0 11.2 10.7 0.8 8.8
# 2027 plain 10.7 11.0 10.7 11.2 11.5 11.3 11.5 11.0 11.2 0.0 0.0
# 2027 11/22 10.7 9.3 10.7 1.4 11.5 11.3 11.5 11.0 11.2 1.6 9.9
# plain shares: lowest 10.68, highest 11.50, ratio 1.077
# 2027 shares are the 2026 shares moved up by one: True
#
# Reading: without master numbers the nine personal years hold
# nearly equal shares of people, the same shares every year, only
# relabelled. With 11 and 22 kept, 22 takes most of the people
# who would be in year 4, while 11 takes only a few from year 2.不保留大师数时,九个流年的比例在 10.68 到 11.50 之间,比值 1.077,而且每年都是同样的九个比例,只是标签轮转:程序验证了 2027 年的比例恰好是 2026 年的比例向上移一位。保留 11 与 22 时,22 拿走了本应处于流年 4 的大部分人,11 只从流年 2 拿走少数人。所以「今年是很多人的流年 4」这样的说法,取决于软件是否保留大师数,而不取决于年份。
小结
公历月份的长短几乎不在余数上留下偏倚:纯数字根的生命灵数在多年的出生人群中均匀到小数点后一位,一年之内在 10.7 到 11.5 之间。偏倚来自大师数的约定:它让 2、4、6 成为最少见的数,并让 1990 年代出生的人中没有一个人的生命灵数是 2。流年数的人群比例每年相同,只是标签轮转。
参考文献
Fan, Y. (2026). Modular Numerology, Character-Stroke Hashes and Residue-Class Distributions: Digital-Root Arithmetic, Kangxi Radical Restorations, Modulo-80 Grid Mappings and Empirical Base-Rate Concentration [姓名笔画、生命灵数与手机号码的同余映射和分布论] (Version 1.0.0). 5D Library. https://doi.org/10.5281/zenodo.24090178 · Licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0).