What does Studio Tariffs, Prepaid Float and Onboarding conclude, in numbers, about “Breakage Bounds and the Cost of Spending the Float”?

Answer: Chapter 4 shows that breakage estimated after twelve months is 0.2358 against a true 0.3409, and that keeping 0.7 of the liability in cash fails on 0.2225 of paths.

Scope of the numbers: Its clients, demands, rates and parameters are illustrative models; all calculations are generated by its programs; the book is not legal, financial, accounting or tax advice. Chapter 7 reads the New Book of Tang on flying cash, the Song office that let official buildings and the theory of two-sided markets: Little's law gives 17500 against a simulated 17554.

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: 0.2358, 0.3409

true breakage: 0.3409
monthspassesDphbreakage95% interval
66000.200.300.00000.20000.1300 to 0.2600
1212000.140.280.01250.23580.1803 to 0.2850
2424000.160.300.02500.32180.2996 to 0.3380
Show the full program
# Chapter 4. p07: How soon can breakage be known? A pass holds 5 visits.
# True model: 0.15 of buyers never come (D); the others visit in a month
# with chance p = 0.30 and, after each month, lapse for good with chance
# h = 0.03. 100 passes are sold each month. After W months of sales the
# platform fits D, p, h by maximum likelihood to every pass's history and
# reads off breakage (credit never used) with a 95% profile interval
# (log-likelihood within 1.92 of the best). Seed 113.
import math
import random

TRUE, CAP, PER = (0.15, 0.30, 0.03), 5, 100

def breakage(d, p, h):
    """Expected share of the 5 visits never used."""
    unused = 0.0
    for m in range(1, 600):
        pm = (1 - h) ** (m - 1) * h
        for x in range(CAP):
            unused += pm * math.comb(m, x) * p ** x * (1 - p) ** (m - x) * \
                (CAP - x) if x <= m else 0.0
    return d + (1 - d) * unused / CAP

def simulate(rng, window):
    hist = {}
    for month in range(window):
        n = window - month                    # months observed
        for _ in range(PER):
            dormant = rng.random() < TRUE[0]
            x, last, alive = 0, 0, True
            for t in range(1, n + 1):
                if dormant or not alive or x == CAP:
                    break
                if rng.random() < TRUE[1]:
                    x, last = x + 1, t
                if rng.random() < TRUE[2]:
                    alive = False
            key = (n, x, last)
            hist[key] = hist.get(key, 0) + 1
    return hist

def loglik(hist, d, p, h):
    ll = []
    for (n, x, last), c in hist.items():
        if x == CAP:
            l = (1 - d) * p ** x * (1 - p) ** (last - x) * \
                (1 - h) ** (last - 1)
        else:
            l = p ** x * (1 - p) ** (n - x) * (1 - h) ** n
            for i in range(max(last, 1), n):
                l += p ** x * (1 - p) ** (i - x) * (1 - h) ** (i - 1) * h
            l = (1 - d) * l + (d if x == 0 else 0.0)
        ll.append(c * math.log(l))
    return math.fsum(ll)

if __name__ == "__main__":
    print("true breakage: {:.4f}".format(breakage(*TRUE)))
    print("months  passes   D     p     h     breakage   95% interval")
    grid = [(d / 100, p / 100, h / 400) for d in range(5, 31)
            for p in range(20, 41, 2) for h in range(0, 25)]
    memo = {}
    for window in (6, 12, 24):
        hist = simulate(random.Random(113), window)
        scored = [(loglik(hist, *g), g) for g in grid]
        top, g = max(scored)
        near = []
        for s, q in scored:
            if s >= top - 1.92:
                if q not in memo:
                    memo[q] = breakage(*q)
                near.append(memo[q])
        print("{:6d}  {:6d}  {:.2f}  {:.2f}  {:.4f}  {:8.4f}   {:.4f} to "
              "{:.4f}".format(window, window * PER, g[0], g[1], g[2],
                              breakage(*g), min(near), max(near)))

Single source and citation

Fan, Y. (2026). Unmanned Shared-Space Revenue Management, Cross-Venue Prepaid Float and Zero-Take-Rate Trojan Onboarding (Version 1.0.0), Chapter 4: Breakage Bounds and the Cost of Spending the Float. 5D Library. DOI pending

DOI pending · monograph page goes live after publication · 5DL-B-113 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026113,
  author    = {Fan, Yafu},
  title     = {Unmanned Shared-Space Revenue Management, Cross-Venue Prepaid Float and Zero-Take-Rate Trojan Onboarding},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

Other chapter answers in this book