What does Diagnostic Paywalls and Unlock Pricing conclude, in numbers, about “Thirty Sites as One Markov Chain”?

Answer: Chapter 4 solves thirty sites as one Markov chain: routing within themes raises purchases by 1.4087, and four links a page is best.

Scope of the numbers: Its users, rates and parameters are illustrative models; all calculations are generated by its programs; the book is not legal, financial or consumer-protection advice. Marked boxes separate published evidence, runnable Python calculations (16 programs), classical sources and the author's hypotheses.

中文版

Data table

The numbers below are printed verbatim by the chapter's Python program and can be reproduced by running it. standard-library Python onlyprinted values that match the answer: 1.4087

policybuys per 1000 arrivalspages per visitvs none
none30.0001.00001.0000
uniform35.5441.42861.1848
8040.6371.42861.3546
9542.2601.42861.4087
Show the full program
# Chapter 4. p07: Thirty sites as one Markov chain. Thirty diagnostic
# sites form 6 themes of 5. A visitor arrives at a site and keeps the
# theme of that site as her interest. On each page she buys (and leaves)
# with chance b_j r, where b_j runs from 0.010 to 0.030 and r is 1.5 on a
# site of her theme and 0.5 elsewhere; she follows a cross-link with
# chance 0.30, else leaves. Policies route a click to another site: none,
# uniform, 80% or 95% within the theme. We solve for purchases per 1000
# arrivals and pages per visit by iterating the chain to convergence.
import math

N, T = 30, 6
B = [0.010 + 0.020 * ((7 * j) % N) / (N - 1) for j in range(N)]

def theme(j):
    return j // 5

def route(policy, j):
    """Probabilities of the next site after a click on site j."""
    p = [0.0] * N
    if policy == "none":
        return p
    same = [k for k in range(N) if theme(k) == theme(j) and k != j]
    other = [k for k in range(N) if theme(k) != theme(j)]
    w_same = {"uniform": len(same) / (N - 1), "80": 0.80, "95": 0.95}[policy]
    for k in same:
        p[k] = w_same / len(same)
    for k in other:
        p[k] = (1 - w_same) / len(other)
    return p

def solve(policy, g):
    buy = [B[j] * (1.5 if theme(j) == g else 0.5) for j in range(N)]
    click = 0.30 if policy != "none" else 0.0
    P = [route(policy, j) for j in range(N)]
    v, n = [0.0] * N, [0.0] * N
    for _ in range(80):
        v = [buy[j] + click * math.fsum(P[j][k] * v[k] for k in range(N))
             for j in range(N)]
        n = [1 + click * math.fsum(P[j][k] * n[k] for k in range(N))
             for j in range(N)]
    return v, n

if __name__ == "__main__":
    print("policy    buys per 1000 arrivals   pages per visit   vs none")
    base = None
    for pol in ("none", "uniform", "80", "95"):
        bs, ps = [], []
        for i in range(N):
            v, n = solve(pol, theme(i))
            bs.append(v[i])
            ps.append(n[i])
        b = 1000 * math.fsum(bs) / N
        base = base or b
        print("{:8s}  {:22.3f}   {:15.4f}   {:7.4f}".format(
            pol, b, math.fsum(ps) / N, b / base))

Single source and citation

Fan, Y. (2026). Algorithmic Diagnostic Paywalls, Curiosity-Gap Calibration and Tiered Report Unlock Economics (Version 1.0.0), Chapter 4: Thirty Sites as One Markov Chain. 5D Library. DOI pending

DOI pending · monograph page goes live after publication · 5DL-B-114 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026114,
  author    = {Fan, Yafu},
  title     = {Algorithmic Diagnostic Paywalls, Curiosity-Gap Calibration and Tiered Report Unlock Economics},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

Other chapter answers in this book