关于「实证闭环:合参权重降权提议与预注册检验」,《十二体系的同源证据降权》给出的定量结论是什么?

结论:第八章给出合参系统的只读降权提议、蒸馏小工具「同源证据降权器」与预注册检验方案:示范中朴素概率 0.917 经降权为 0.750。书中以标注框区分已发表的证据、可运行的 Python 计算(18 段程序)、原典与作者的假说。

数据口径:本书只分析符号体系的结构、信息量与概率,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。 第六章用数据处理不等式说明,示意模型中十二套体系关于结果的信息恰为出生日本身的 0.5769 毫比特,一点不多;事后对上几乎必然,每句陈述只有 0.05 的机会对上时,十二套体系至少一句对上的概率是 0.9539。

English version

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: 0.750, 0.917

naive: 12 systems x 0.2 = log-odds 2.40, P 0.917
within-block rho 1.0: total weight 5.00, log-odds 1.000, P 0.731
within-block rho 0.8: total weight 5.48, log-odds 1.097, P 0.750
within-block rho 0.5: total weight 6.60, log-odds 1.320, P 0.789
detail at rho 0.8
  block   size  weight each  contribution
  solar      3        0.385         0.231
  year       3        0.385         0.231
  day        4        0.294         0.235
  number     1        1.000         0.200
  hour       1        1.000         0.200
status: PROPOSAL-ONLY; read-only; synthesis owner decides
查看完整验算程序
# Chapter 8. p81: the read-only down-weighting proposal for a synthesis.
# Systems are grouped into the shared-cycle blocks of p52. Inside a block
# of m systems with within-block correlation rho, each log-odds is scaled
# by 1 / (1 + (m - 1) rho), the exact Gaussian weight of p41; blocks are
# then added as if orthogonal. Every system here reports the same
# illustrative log-odds of +0.2. The output is a proposal for the
# synthesis owner to accept or reject, not a verdict on any person.
import math

BLOCKS = [("solar", ["TROP", "SIDR", "MONB"]),
          ("year", ["ZODI", "NSTAR", "NAYIN"]),
          ("day", ["DAYP", "MANS", "WEEK", "TZOL"]),
          ("number", ["LIFE"]), ("hour", ["HOUR"])]
L = 0.2


def combine(rho):
    parts = []
    for name, members in BLOCKS:
        m = len(members)
        w = 1 / (1 + (m - 1) * rho)
        parts.append((name, m, w, math.fsum(w * L for _ in members)))
    return parts


def prob(lo):
    return 1 / (1 + math.exp(-lo))


naive = math.fsum(L for _, mem in BLOCKS for _ in mem)
print("naive: 12 systems x %.1f = log-odds %.2f, P %.3f"
      % (L, naive, prob(naive)))
for rho in (1.0, 0.8, 0.5):
    parts = combine(rho)
    tot = math.fsum(p[3] for p in parts)
    wt = math.fsum(p[1] * p[2] for p in parts)
    print("within-block rho %.1f: total weight %.2f, log-odds %.3f, P %.3f"
          % (rho, wt, tot, prob(tot)))
print("detail at rho 0.8")
print("  block   size  weight each  contribution")
for name, m, w, c in combine(0.8):
    print("  %-7s %4d %12.3f %13.3f" % (name, m, w, c))
print("status: PROPOSAL-ONLY; read-only; synthesis owner decides")

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《十二命理体系独立证据降权、同源生辰伪共振剥离与跨东西方符号系统正交合参潜变量母典》第 8 章「实证闭环:合参权重降权提议与预注册检验」(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-283 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026283,
  author    = {Fan, Yafu},
  title     = {Down-Weighting Non-Independent Evidence across Twelve Symbolic Systems: Spurious Resonance from a Shared Birth Input and Orthogonal Latent-Variable Synthesis},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论