关于「作品与辩护:有成本信号、累赘原理与廉价话语」,《零解释税与作品引力》给出的定量结论是什么?

结论:第五章比较作品与辩护两类信号:看 4 件可核验样本的判断准确率是 0.770,用自述达到同样准确率需要 66 句。

数据口径:书中参数一律为示意,全部计算由书中程序生成;本书是非个人化经营建议,不是心理咨询或临床判断,不作任何收益承诺,「显化」只指把构想做成可交付、可核验的作品。 第三章求出解释的盈亏阈值,示意参数下为 0.464,只对阈值以上受众解释的净值是 320.0,对所有人解释是负 203.8。

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: 0.770, 66

A. accuracy / matched found / bits, by number of items k
kartifactstalk
10.700 / 0.000 / 0.0560.700 / 0.000 / 0.003
20.735 / 0.490 / 0.1070.700 / 0.000 / 0.007
40.770 / 0.652 / 0.1990.700 / 0.000 / 0.014
80.831 / 0.552 / 0.3480.700 / 0.000 / 0.027
160.907 / 0.825 / 0.5490.704 / 0.185 / 0.053
320.965 / 0.931 / 0.7460.725 / 0.367 / 0.102
B. guessing 'unmatched' for all gives accuracy 0.700
C. talk items needed to match 4 artifacts: 66
查看完整验算程序
# Chapter 5. p53: how well a reader sorts creators from k items.
# Prior share of creators matched to this reader: 0.3. A verifiable
# sample (artifact) looks matched with probability 0.70 if the creator
# is matched and 0.40 if not; a self-description (talk) "sounds matched"
# with probability 0.90 vs 0.85, since anyone can say it. The reader
# applies the Bayes rule to the count of positive items; we compute the
# exact accuracy and the share of matched creators found. Illustrative.
import math

PRIOR = 0.3


def pmf(k, x, p):
    return math.comb(k, x) * p ** x * (1 - p) ** (k - x)


def sort_quality(k, p1, p0):
    acc, found = [], []
    for x in range(k + 1):
        a = PRIOR * pmf(k, x, p1)
        b = (1 - PRIOR) * pmf(k, x, p0)
        if a > b:
            acc.append(a)
            found.append(a)
        else:
            acc.append(b)
    return math.fsum(acc), math.fsum(found) / PRIOR


def bits(k, p1, p0):
    terms = []
    for x in range(k + 1):
        a = PRIOR * pmf(k, x, p1)
        b = (1 - PRIOR) * pmf(k, x, p0)
        px = a + b
        for joint, prior in ((a, PRIOR), (b, 1 - PRIOR)):
            if joint > 0:
                terms.append(joint * math.log2(joint / (px * prior)))
    return math.fsum(terms)


if __name__ == "__main__":
    print("A. accuracy / matched found / bits, by number of items k")
    print("   k     artifacts               talk")
    for k in (1, 2, 4, 8, 16, 32):
        a1, f1 = sort_quality(k, 0.70, 0.40)
        a2, f2 = sort_quality(k, 0.90, 0.85)
        cell = "{:.3f} / {:.3f} / {:.3f}"
        line = "   {:>2}    " + cell + "   " + cell
        print(line.format(k, a1, f1, bits(k, 0.70, 0.40),
                          a2, f2, bits(k, 0.90, 0.85)))
    print("B. guessing 'unmatched' for all gives accuracy {:.3f}".format(
        1 - PRIOR))
    need = next(k for k in range(1, 400)
                if sort_quality(k, 0.90, 0.85)[0] >=
                sort_quality(4, 0.70, 0.40)[0])
    print("C. talk items needed to match 4 artifacts: {}".format(need))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《零解释税生存动力学:放弃低维认同求证、全功率自我显化与作品引力场母典》第 5 章「作品与辩护:有成本信号、累赘原理与廉价话语」(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-432 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026432,
  author    = {Fan, Yafu},
  title     = {Zero Explanation-Tax Dynamics: Abandoning Validation-Seeking, Full-Power Reality Rendering and Proof-by-Artifact Gravity},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论