5D Library › Prediction Market Microstructure, Bregman Information Projection and Cross-Book Arbitrage Bounds · 中文
Chapter 1. A Price Is Not a Probability: Belief Dispersion, Risk Attitudes and the Favourite-Longshot Bias
A popular reading
A prediction-market contract pays 1 if an event happens and 0 if not. Hence a popular reading: a contract at 0.20 means the market gives the event a 20 % chance. Wolfers and Zitzewitz’s survey made the reading widely known (Wolfers and Zitzewitz 2004), and in many settings it works well. But it is a conclusion, and it needs premises. This chapter finds the premises and shows which way the price leans when they fail.
First, the problem. A market has many traders, each holding a belief about the event. It has one price. Compressing a distribution into one number requires choosing a statistic: the mean, the median, or some other quantile. Which one the price picks depends on the traders’ budgets, their attitudes to risk and the shape of the belief distribution, not on what we would like it to be.
Manski’s result: the price is a quantile
Manski answered the question with the simplest possible model (Manski 2006): traders are risk-neutral with equal budgets; those whose belief exceeds the price spend everything on YES, the rest everything on NO.
Calculation
The clearing condition. Let be the distribution function of beliefs and the price. Each YES buyer gets contracts for a budget of 1, each NO buyer . YES and NO contracts must match: The price is the point at which the share of traders believing more than it equals the price itself. Manski further shows that, however beliefs are distributed, their mean lies strictly inside — and that is all the price says about the mean.
Calculation
# Chapter 1. p01: Manski's result: a market price is not the mean belief.
# Risk-neutral traders hold equal budgets and a belief q that a binary
# event happens. A trader with q above the price spends the budget on YES
# contracts (each pays 1 if the event happens), one below it on NO.
# Clearing (YES bought = NO bought) gives F(price) = 1 - price: the price
# is the point where the SHARE of traders above it equals the price.
# Manski (2006): the mean belief lies strictly inside
# (price^2, 2*price - price^2), whatever the belief distribution.
# Section B gives budgets that rise with belief: then the price is the
# point where the WEALTH share above it equals the price.
import bisect
import random
N = 20000
def clear(beliefs, budgets=None):
pairs = sorted(zip(beliefs, budgets or [1.0] * len(beliefs)))
qs = [q for q, _ in pairs]
cum = [0.0]
for _, w in pairs:
cum.append(cum[-1] + w)
total = cum[-1]
lo, hi = 1e-9, 1 - 1e-9
for _ in range(60):
mid = (lo + hi) / 2
above = (total - cum[bisect.bisect_left(qs, mid)]) / total
lo, hi = (mid, hi) if above > mid else (lo, mid)
return (lo + hi) / 2
def draw(spec, rng):
if spec[0] == "beta":
return [rng.betavariate(spec[1], spec[2]) for _ in range(N)]
out = []
for _ in range(N):
a, b = spec[1] if rng.random() < spec[3] else spec[2]
out.append(rng.betavariate(a, b))
return out
if __name__ == "__main__":
rng = random.Random(59001)
pops = [("symmetric Beta(4,4)", ("beta", 4, 4)),
("pessimist Beta(2,8)", ("beta", 2, 8)),
("optimist Beta(8,2)", ("beta", 8, 2)),
("longshot Beta(1.5,13.5)", ("beta", 1.5, 13.5)),
("split 70/30 bimodal", ("mix", (2, 18), (18, 2), 0.7))]
print("A. equal budgets")
print(" population mean median price (p^2, 2p-p^2)")
kept = {}
for name, spec in pops:
q = draw(spec, rng)
kept[name] = q
p = clear(q)
mean, med = sum(q) / N, sorted(q)[N // 2]
ok = p * p < mean < 2 * p - p * p
print(" {:<26}{:5.3f} {:6.3f} {:5.3f} ({:5.3f}, {:5.3f}) {}"
.format(name, mean, med, p, p * p, 2 * p - p * p,
"inside" if ok else "OUTSIDE"))
print("\nB. Beta(1.5,13.5) beliefs; budget = q ** k (richer if bolder)")
q = kept["longshot Beta(1.5,13.5)"]
mean = sum(q) / N
for k in (0.0, 0.5, 1.0, 2.0):
w = [x ** k for x in q]
wmean = sum(a * b for a, b in zip(q, w)) / sum(w)
p = clear(q, w)
print(" k = {:3.1f} mean {:5.3f} wealth-weighted mean {:5.3f}"
" price {:5.3f}".format(k, mean, wmean, p))
print("\nReading: the price sits where the share (or wealth share) of")
print("traders above it equals it - a quantile, not the mean. With the")
print("same mean belief, skew and budgets move the price; only Manski's")
print("interval constrains the mean.")
# A. equal budgets
# population mean median price (p^2, 2p-p^2)
# symmetric Beta(4,4) 0.501 0.501 0.501 (0.251, 0.751) inside
# pessimist Beta(2,8) 0.201 0.181 0.267 (0.071, 0.462) inside
# optimist Beta(8,2) 0.800 0.822 0.735 (0.541, 0.930) inside
# longshot Beta(1.5,13.5) 0.100 0.083 0.167 (0.028, 0.306) inside
# split 70/30 bimodal 0.341 0.127 0.307 (0.094, 0.520) inside
#
# B. Beta(1.5,13.5) beliefs; budget = q ** k (richer if bolder)
# k = 0.0 mean 0.100 wealth-weighted mean 0.100 price 0.167
# k = 0.5 mean 0.100 wealth-weighted mean 0.128 price 0.193
# k = 1.0 mean 0.100 wealth-weighted mean 0.154 price 0.216
# k = 2.0 mean 0.100 wealth-weighted mean 0.202 price 0.257
#
# Reading: the price sits where the share (or wealth share) of
# traders above it equals it - a quantile, not the mean. With the
# same mean belief, skew and budgets move the price; only Manski's
# interval constrains the mean.The five populations of Part A show the gap between a quantile and a mean. With symmetric beliefs, price, mean and median coincide at 0.501. With skewed beliefs they do not: a “longshot” population whose average belief is 0.100, with median 0.083, trades at 0.167 — two thirds above the mean. The reason lies in the clearing condition itself: at a low price a NO contract costs nearly 1 while a YES contract costs little, so the budgets of a few optimists buy a great many YES contracts. The same mechanism pulls an optimistic population (mean 0.800) down to 0.735. The bimodal population has mean 0.341 and median 0.127 and trades at 0.307, different from both. All five means fall inside Manski’s interval, but the interval is wide: for the longshot population it is (0.028, 0.306).
Part B lets budgets rise with belief: the bolder, the richer. The distribution of beliefs is unchanged, yet the price climbs from 0.167 to 0.257. The price now follows a quantile of the wealth distribution: the share of wealth above the price equals the price.
Log utility: the one case where price equals the mean
Wolfers and Zitzewitz give the result at the other end: if every trader has log utility, the price is exactly the wealth-weighted mean belief (Wolfers and Zitzewitz 2006). Gjerstad studied how risk aversion moves the price away from that mean (Gjerstad 2005). The second program checks constant-relative-risk-aversion (CRRA) utility case by case.
Calculation
Demand of a CRRA trader. A trader with belief , wealth and risk aversion holds YES contracts (negative for NO). The first-order condition for maximum expected utility gives and the price sets .
Calculation
# Chapter 1. p02: Wolfers-Zitzewitz: with log utility the price IS the
# wealth-weighted mean belief; other risk attitudes move it away.
# Trader i has belief q, wealth w and CRRA risk aversion g. Holding n YES
# contracts bought at price p, terminal wealth is w + n(1 - p) if the
# event happens and w - n p if not. Maximising expected utility gives
# R = (q (1 - p) / ((1 - q) p)) ** (1 / g)
# n = w (R - 1) / ((1 - p) + R p)
# (n < 0 is a NO position). The price clears sum(n) = 0.
import math
import random
N = 4000
def demand(q, w, g, p):
r = (q * (1 - p) / ((1 - q) * p)) ** (1.0 / g)
return w * (r - 1) / ((1 - p) + r * p)
def clear(traders):
lo, hi = 1e-6, 1 - 1e-6
for _ in range(60):
mid = (lo + hi) / 2
excess = math.fsum(demand(q, w, g, mid) for q, w, g in traders)
lo, hi = (mid, hi) if excess > 0 else (lo, mid)
return (lo + hi) / 2
if __name__ == "__main__":
rng = random.Random(59002)
q = [rng.betavariate(1.5, 13.5) for _ in range(N)]
w = [math.exp(rng.gauss(0, 1)) for _ in range(N)]
mean = math.fsum(q) / N
wmean = math.fsum(a * b for a, b in zip(q, w)) / math.fsum(w)
med = sorted(q)[N // 2]
print("beliefs Beta(1.5,13.5): mean {:.4f}, wealth-weighted {:.4f},"
" median {:.4f}".format(mean, wmean, med))
print("\n risk aversion g price price - wtd mean")
cases = [("all g = 1 (log)", lambda x: 1.0),
("all g = 3", lambda x: 3.0),
("all g = 0.5 (risk-loving)", lambda x: 0.5),
("bold g=0.5, timid g=4", lambda x: 0.5 if x > med else 4.0),
("bold g=4, timid g=0.5", lambda x: 4.0 if x > med else 0.5)]
for name, gf in cases:
p = clear([(a, b, gf(a)) for a, b in zip(q, w)])
d = round(p - wmean, 4) + 0.0 # + 0.0 turns -0.0 into 0.0
print(" {:<26} {:6.4f} {:+.4f}".format(name, p, d))
print("\nReading: log utility returns the wealth-weighted mean exactly.")
print("Strong uniform risk aversion pulls the price toward the median")
print("belief, risk-loving toward the bold; when risk attitude goes with")
print("belief the price moves furthest, in either direction. A price")
print("reads as a probability only under an assumption about who trades")
print("and how they weigh risk.")
# beliefs Beta(1.5,13.5): mean 0.0984, wealth-weighted 0.0989, median 0.0819
#
# risk aversion g price price - wtd mean
# all g = 1 (log) 0.0989 +0.0000
# all g = 3 0.0823 -0.0167
# all g = 0.5 (risk-loving) 0.1171 +0.0182
# bold g=0.5, timid g=4 0.1444 +0.0454
# bold g=4, timid g=0.5 0.0560 -0.0429
#
# Reading: log utility returns the wealth-weighted mean exactly.
# Strong uniform risk aversion pulls the price toward the median
# belief, risk-loving toward the bold; when risk attitude goes with
# belief the price moves furthest, in either direction. A price
# reads as a probability only under an assumption about who trades
# and how they weigh risk.Four thousand traders have beliefs distributed like p01’s longshot population and log-normal wealth. Under log utility the price is 0.0989, exactly the wealth-weighted mean, which checks the program itself. When everyone has risk aversion 3 the price falls to 0.0823, almost exactly the median of 0.0819; when everyone is risk-loving () it rises to 0.1171. The largest departures come when risk attitude goes with belief: if the bold are also unafraid of risk and the cautious also afraid of it, the price is 0.1444, 0.0454 above the wealth-weighted mean; the other way round it is 0.0429 below.
So “price = probability” needs more than traders who are unbiased on the whole. It needs a specific assumption about who trades and how they weigh risk, and the price alone usually cannot test it.
The favourite-longshot bias, with no one loving risk
In horse-race betting the odds on outsiders have long been too generous and those on favourites too mean: a unit bet on longshots returns less on average than one on favourites. This favourite-longshot bias was first recorded by Griffith (Griffith 1949), measured systematically on racetrack data by Ali (Ali 1977) and surveyed by Thaler and Ziemba (Thaler and Ziemba 1988). The two usual explanations are that bettors love risk or that they overweight small probabilities (Snowberg and Wolfers 2010).
The third program tests a third possibility: no preference and no cognitive bias at all, only disagreement. Each trader sees the true probability plus unbiased noise.
Calculation
# Chapter 1. p03: The favourite-longshot bias from belief dispersion alone.
# 12,000 binary contracts. Each has a true probability p (uniform on the
# log-odds scale between -4 and +4). 200 traders per contract see p with
# unbiased noise on the log-odds scale (sd 1.0), so no one is biased on
# average and no one loves risk. Two clearing rules are compared:
# M Manski: risk-neutral, equal budgets (the quantile price of p01)
# L log utility, equal wealth (price = mean belief, as in p02)
# Contracts are grouped by price; each group's mean price is compared
# with its mean true probability and with simulated outcomes.
import bisect
import math
import random
K, T, SD = 12000, 200, 1.0
EDGES = (0.0, 0.05, 0.10, 0.20, 0.40, 0.60, 0.80, 0.90, 0.95, 1.0)
def sig(x):
return 1 / (1 + math.exp(-x))
def manski(qs):
qs = sorted(qs)
lo, hi = 1e-9, 1 - 1e-9
for _ in range(50):
mid = (lo + hi) / 2
above = 1 - bisect.bisect_left(qs, mid) / len(qs)
lo, hi = (mid, hi) if above > mid else (lo, mid)
return (lo + hi) / 2
def table(rows, label):
print("\n{} price band n mean price true p observed"
.format(label))
for lo, hi in zip(EDGES, EDGES[1:]):
g = [r for r in rows if lo <= r[0] < hi]
if not g:
continue
n = len(g)
mp = sum(r[0] for r in g) / n
tp = sum(r[1] for r in g) / n
ob = sum(r[2] for r in g) / n
print(" [{:.2f}, {:.2f}) {:5d} {:6.3f} {:6.3f} {:6.3f}"
.format(lo, hi, n, mp, tp, ob))
if __name__ == "__main__":
rng = random.Random(59003)
rows_m, rows_l = [], []
for _ in range(K):
x = rng.uniform(-4, 4)
p = sig(x)
qs = [sig(x + rng.gauss(0, SD)) for _ in range(T)]
y = 1 if rng.random() < p else 0
rows_m.append((manski(qs), p, y))
rows_l.append((sum(qs) / T, p, y))
table(rows_m, "M")
table(rows_l, "L")
print("\nReading: with unbiased but dispersed beliefs, both rules")
print("overprice longshots and underprice favourites - the bias needs")
print("neither risk-love nor misperception, only disagreement plus a")
print("price bounded by 0 and 1.")
#
# M price band n mean price true p observed
# [0.05, 0.10) 802 0.084 0.024 0.022
# [0.10, 0.20) 1905 0.145 0.058 0.065
# [0.20, 0.40) 2379 0.296 0.202 0.206
# [0.40, 0.60) 2039 0.500 0.499 0.502
# [0.60, 0.80) 2287 0.707 0.800 0.813
# [0.80, 0.90) 1798 0.854 0.942 0.946
# [0.90, 0.95) 790 0.914 0.976 0.977
#
# L price band n mean price true p observed
# [0.00, 0.05) 971 0.038 0.025 0.025
# [0.05, 0.10) 1212 0.072 0.050 0.049
# [0.10, 0.20) 1369 0.145 0.111 0.129
# [0.20, 0.40) 1799 0.294 0.260 0.260
# [0.40, 0.60) 1502 0.500 0.499 0.503
# [0.60, 0.80) 1708 0.705 0.738 0.752
# [0.80, 0.90) 1335 0.854 0.888 0.894
# [0.90, 0.95) 1178 0.927 0.950 0.954
# [0.95, 1.00) 926 0.962 0.975 0.977
#
# Reading: with unbiased but dispersed beliefs, both rules
# overprice longshots and underprice favourites - the bias needs
# neither risk-love nor misperception, only disagreement plus a
# price bounded by 0 and 1.Both clearing rules produce the favourite-longshot bias. Under Manski’s rule, contracts priced between 0.05 and 0.10 average 0.084, while their true probability is only 0.024 and the simulated outcomes occur at 0.022; contracts priced between 0.90 and 0.95 average 0.914 against a true 0.976. Under log utility the bias is smaller but in the same direction: longshots at 0.038 have a true probability of 0.025, favourites at 0.962 one of 0.975; the middle band (0.500 against 0.499) is almost unbiased.
The mechanism is simple. Prices are confined between 0 and 1, while the noise is symmetric in log-odds. When the true probability is small, noise pushes some beliefs up a long way and others down only a little — that side is already close to 0. Averaging those beliefs (log utility) or taking Manski’s quantile pulls the result toward 0.5 either way.
Note
What this shows and what it does not. It shows that the favourite-longshot bias can arise entirely from dispersed beliefs and bounded prices, so seeing the bias in data does not prove that traders love risk or overweight small chances. It does not show which mechanism operates in real markets; Snowberg and Wolfers separated the risk-love and misperception explanations using combinations of pari-mutuel bets (Snowberg and Wolfers 2010), and this book does not repeat that work.
Hypothesis · not established
The author’s hypothesis (not yet tested): on one platform, the more traders disagree about a contract (measured, say, by the dispersion of independent forecasts of the same event), the more its longshots are overpriced; for low-disagreement contracts the favourite-longshot bias is near zero. What would show it wrong: after grouping by disagreement, the bias is the same in every group, or larger where disagreement is low.
Summary
A prediction-market price is a statistic of the belief distribution, and the trading rules choose the statistic. With risk-neutral traders and equal budgets the price is the quantile at which the share of traders above it equals it, and the mean belief can only be bounded within ; with log utility the price is exactly the wealth-weighted mean; with other attitudes to risk, especially when risk attitude goes with belief, the price can depart in either direction. As long as beliefs disagree and prices are bounded, longshots are overpriced and favourites underpriced without anyone loving risk. Before reading a price, ask who set it.
References
Fan, Y. (2026). Prediction Market Microstructure, Bregman Information Projection and Cross-Book Arbitrage Bounds: Logarithmic Market Scoring Rules, Favorite-Longshot Calibration Bias, Lock-Up Opportunity Cost and Frank-Wolfe Combinatorial Pricing [预测市场微观结构、布雷格曼信息投影与跨盘口无套利边界] (Version 1.0.0). 5D Library. https://doi.org/10.5281/zenodo.24090166 · Licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0).