Risk-Adjusted Mispricing Test
Bet only when your estimated odds beat the price by enough to justify the loss
- Difficulty
- Advanced
- Time to result
- ~ongoing to results
- Steps
- 6
- Confidence
- 95%
Galloway explains why he planned to bet on Kamala Harris even while acknowledging that a statistician might narrowly favour Donald Trump. Prediction-market prices implied a much lower chance for Harris than the close polling suggested to him. His decision rule separates the forecast from the wager: estimate probability from independent evidence, compare it with the price-implied probability, and act only when the payoff appears to overcompensate for the risk. He also proposes a source of mispricing, arguing that prediction-market users skewed younger, male, and pro-Trump. The method still requires a loss limit and a bias check. Galloway openly says he wants Harris to win and may be finding supportive reasons. The episode records a proposed wager, not a validated return, and no apparent edge makes a large stake safe.
Origin
On The Diary of a CEO, Galloway applies the test to a planned Harris prediction-market wager while repeatedly acknowledging uncertainty and confirmation bias.
Core principles
- 01Separate an outcome's likelihood from the market price offered on it
- 02A likely loser can still be an attractive bet when the payoff is sufficiently mispriced
- 03Use independent evidence rather than the preferences of the market's participants
- 04Positive expected value does not remove the possibility of a total loss
How to run it
- 1
Read the market price
Convert the quoted price or payout into the probability the market appears to imply. Record fees and settlement conditions before comparing anything.
Pro tip Write the implied probability down so enthusiasm cannot silently change it.
Watch out A headline payout can obscure fees, liquidity, or settlement risk.
- 2
Build an independent estimate
Use relevant evidence outside the market price to estimate the outcome's probability. Prefer multiple credible signals over one vivid story.
Pro tip Galloway compares prediction-market odds with high-quality polls.
Watch out Your estimate can be wrong even when it feels well supported.
- 3
Explain the disagreement
Identify why market participants might price the event differently from your evidence. A plausible mechanism strengthens the case but does not prove mispricing.
Pro tip Check whether participant demographics or incentives could skew the market.
Watch out Disagreement alone is not an edge; the market may know something you do not.
- 4
Test the expected trade-off
Compare potential gain, probability of winning, and full downside. Reject the opportunity if a modest estimation error removes the apparent advantage.
Pro tip Run the comparison with a less favourable probability than your central estimate.
Watch out Expected value is an average across possible outcomes, not a promised result on one wager.
- 5
Audit your motive
Ask whether identity, hope, fear, or recent anecdotes are inflating your probability estimate. Seek evidence that would make you decline the bet.
Pro tip State your preferred outcome separately from your forecast.
Watch out Galloway explicitly identifies confirmation bias in his own election reasoning.
- 6
Size for survival
Risk only an amount whose complete loss does not threaten essential goals or trigger further chasing. If no safe size exists, do not act.
Watch out The transcript gives Galloway's proposed stake but does not establish that it is suitable for anyone else.
In the wild
Galloway says prediction-market pricing was roughly 62 to 38 in Trump's favour while high-quality polls showed a toss-up with a slight Trump advantage. He therefore viewed a Harris payout of roughly 2.8 or 2.9 to one as attractive, despite admitting he did not know the result and preferred Harris politically.
→ The example shows how a person can favour one candidate as the likelier winner yet view the other candidate as the better-priced wager; the transcript does not report the wager's execution or result.
A supplier estimates a 45% chance of winning a tender after reviewing comparable bids, while its internal market behaves as if the chance were 20%. Before spending on the proposal, the team checks why the estimates differ, models the full bid cost, lowers its estimate in a stress test, and proceeds only if the remaining expected upside justifies an affordable loss.
→ The team makes a bounded allocation decision rather than confusing a favourable price with certainty.
Common mistakes
Confusing a good bet with a likely winner
An underdog can offer attractive odds while still being more likely to lose the single event.
Estimating from preference
Wanting an outcome can make supportive evidence feel stronger, which Galloway acknowledges in his own forecast.
Ignoring ruin
A theoretical edge does not justify risking essential money or a stake the decision-maker cannot absorb.
Is it for you?
Best for
It is best for bounded probabilistic decisions with explicit prices, payoffs, credible evidence, and a loss the decision-maker can afford.
Not ideal for
It is not ideal for essential savings, opaque payoffs, compulsive gambling, or situations where the probability estimate is mostly preference.
From the transcript
“on a risk adjusted basis you're getting free risk adjusted return”
“the odds of her winning are not one and three”
“it's total confirmation bias”
From the episode
Scott Galloway (US ELECTION SPECIAL): “I bet $358,000 That They Win The Election!”, “33% Chance Trump Dies In Office!”, “Men Will Have Less Sex If Trump Wins!”