What does Disaster Recovery and Threshold Shares conclude, in numbers, about “Loss and Collusion: Unequal Holders, Shared Buildings and Acquaintance Graphs”?

Answer: Chapter 3 handles unequal holders, shared buildings and collusion among acquaintances: three-of-five split three and two over two places has a loss probability of 0.012547, worse than one place.

Scope of the numbers: Chapter 6 computes self-rescue after one share goes missing: losing the fastest holder is like raising the threshold by one, and the illustrative rescue window is 263 hours. Marked boxes separate published evidence, runnable Python calculations (19 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: 0.012547

site disaster 0.010, break-in 0.005, piece q 0.05, p 0.02
layoutP(lost)P(leaked)
3 copies, one house0.0101240.063514
3 copies, 3 sites0.0002110.072855
3-of-5, one house0.0111470.005077
3-of-5, sites 3+20.0125470.005369
3-of-5, sites 2+2+10.0043550.000745
3-of-5, 5 sites0.0019230.000149
2-of-5, sites 2+2+10.0002950.014144
2-of-5, 5 sites0.0000600.005897
Show the full program
# Chapter 3. p32: shares that share a building fail together.
# Model: pieces sit at sites. A site disaster (fire, flood, seizure of a
# building) destroys every piece there with probability D per year; a
# site break-in exposes every piece there with probability B. On top,
# each piece is lost with q and leaked with p on its own. Exact by
# enumerating site outcomes, then a binomial over the surviving pieces.
# D = 0.01, B = 0.005, q = 0.05, p = 0.02 are illustrative.
import itertools
import math

D, B, Q, P = 0.01, 0.005, 0.05, 0.02


def tail(n, r, x):
    if x <= 0:
        return 1.0
    terms = [math.comb(n, j) * r ** j * (1 - r) ** (n - j)
             for j in range(x, n + 1)]
    return math.fsum(terms)


def risk(layout, k):
    n = int(math.fsum(layout))
    lo, le = [], []
    for hit in itertools.product((0, 1), repeat=len(layout)):
        w_d = math.prod(D if h else 1 - D for h in hit)
        w_b = math.prod(B if h else 1 - B for h in hit)
        hit_n = int(math.fsum(c for c, h in zip(layout, hit) if h))
        alive = n - hit_n
        # as disasters: fewer than k surviving pieces stay usable
        lo.append(w_d * tail(alive, Q, alive - k + 1))
        # as break-ins: exposed pieces plus independent leaks reach k
        le.append(w_b * tail(alive, P, max(0, k - hit_n)))
    return math.fsum(lo), math.fsum(le)


CASES = [("3 copies, one house", [3], 1),
         ("3 copies, 3 sites", [1, 1, 1], 1),
         ("3-of-5, one house", [5], 3),
         ("3-of-5, sites 3+2", [3, 2], 3),
         ("3-of-5, sites 2+2+1", [2, 2, 1], 3),
         ("3-of-5, 5 sites", [1, 1, 1, 1, 1], 3),
         ("2-of-5, sites 2+2+1", [2, 2, 1], 2),
         ("2-of-5, 5 sites", [1, 1, 1, 1, 1], 2)]
print("site disaster %.3f, break-in %.3f, piece q %.2f, p %.2f"
      % (D, B, Q, P))
print("layout                 P(lost)    P(leaked)")
for name, lay, k in CASES:
    lo, le = risk(lay, k)
    print("%-20s %9.6f  %11.6f" % (name, lo, le))

Single source and citation

Fan, Y. (2026). Disaster Recovery for Digital Assets: Shamir Secret-Sharing Thresholds, Off-Site Incremental Snapshots and Self-Rescue from Single Points of Loss (Version 1.0.0), Chapter 3: Loss and Collusion: Unequal Holders, Shared Buildings and Acquaintance Graphs. 5D Library. DOI pending

DOI pending · monograph page goes live after publication · 5DL-B-272 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026272,
  author    = {Fan, Yafu},
  title     = {Disaster Recovery for Digital Assets: Shamir Secret-Sharing Thresholds, Off-Site Incremental Snapshots and Self-Rescue from Single Points of Loss},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

Other chapter answers in this book