What does Martingales, Stopping Times and Stop-Loss Bounds conclude, in numbers, about “Quitting While Ahead Changes Only the Shape: the Win-Rate Illusion and Selling Winners While Holding Losers”?

Answer: Chapter 4 shows that quitting while ahead changes only the shape, and that selling winners early while riding losers has expectation minus 0.270.

Scope of the numbers: Its parameters are illustrative; all calculations are generated by its programs. It is not individual financial advice, does not encourage gambling of any kind, and analyses doubling only as a counter-example.

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: 0.270

half good (p = 0.55), half bad (p = 0.45), deadline 26
ruleE goodT goodE badT badE mix
hold2.60026.00-2.60026.000.000
take+31.40814.08-1.94719.47-0.270
stop-31.94719.47-1.40814.080.270
+3/-30.8588.58-0.8588.580.000
E mix = 0.05 x (T good - T bad): a rule earns only by keeping
good trials open longer than bad ones
Show the full program
# Chapter 4. p43: Cutting winners early and riding losers is the worst
# pairing. Model: half of all trials are good (weekly success p = 0.55,
# drift +0.1) and half are bad (p = 0.45, drift -0.1); the owner cannot
# tell which. Deadline 26 weeks. Rules:
#   hold      run every trial to the deadline
#   take+3    close at +3 (bank the win), never cut a loss
#   stop-3    close at -3 (cut the loss), let wins run
#   +3/-3     both lines
# Wald I per quality: E[S_T | good] = 0.1 E[T | good], and the bad side
# likewise with -0.1. Exact fractions; parameters are illustrative.
from fractions import Fraction

N = 26
RULES = [("hold", None, None), ("take+3", 3, None),
         ("stop-3", None, 3), ("+3/-3", 3, 3)]


def run(p, a, b):
    q = 1 - p
    live = {0: Fraction(1)}
    es = et = Fraction(0)
    for t in range(1, N + 1):
        new = {}
        for s, pr in live.items():
            new[s + 1] = new.get(s + 1, Fraction(0)) + pr * p
            new[s - 1] = new.get(s - 1, Fraction(0)) + pr * q
        live = {}
        for s in sorted(new):
            hit = (a is not None and s >= a) or (b is not None and s <= -b)
            if hit or t == N:
                es += s * new[s]
                et += t * new[s]
            else:
                live[s] = new[s]
    return es, et


print("half good (p = 0.55), half bad (p = 0.45), deadline %d" % N)
print("%-7s %8s %8s %8s %8s %8s" % (
    "rule", "E good", "T good", "E bad", "T bad", "E mix"))
for name, a, b in RULES:
    eg, tg = run(Fraction(11, 20), a, b)
    eb, tb = run(Fraction(9, 20), a, b)
    mix = (eg + eb) / 2
    print("%-7s %8.3f %8.2f %8.3f %8.2f %8.3f" % (
        name, float(eg), float(tg), float(eb), float(tb), float(mix)))
print("E mix = 0.05 x (T good - T bad): a rule earns only by keeping")
print("good trials open longer than bad ones")

Single source and citation

Fan, Y. (2026). Doob's Martingales and the Optional Stopping Theorem: the Doubling Fallacy and Rigorous Stop-Loss Bounds for Business Trials (Version 1.0.0), Chapter 4: Quitting While Ahead Changes Only the Shape: the Win-Rate Illusion and Selling Winners While Holding Losers. 5D Library. DOI pending

DOI pending · monograph page goes live after publication · 5DL-B-249 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026249,
  author    = {Fan, Yafu},
  title     = {Doob's Martingales and the Optional Stopping Theorem: the Doubling Fallacy and Rigorous Stop-Loss Bounds for Business Trials},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

Other chapter answers in this book