rationality255 words

There are no veterans in Russian Roulette

You explore the paradox between mathematical expected value and human risk aversion through the lens of Russian Roulette. While spreadsheets show an 83% win rate as favorable odds, your body's instinctive rejection reveals a deeper trut…

Five chambers out of six say go. An 83 percent win rate, better than most trades, most startups, most marriages. Any expected-value spreadsheet blesses the bet. Every cell of your body refuses it.

They disagree because they are pricing different games. The spreadsheet prices one pull, averaged across many players. Your body prices you, playing until the game ends, and it knows that against a chamber that ends everything, 83 percent is not a probability. It's a countdown.

Behavioral economics calls the body's vote a bias. Losses land [about twice as heavy as equivalent gains](Loss aversion explains why humans fear failure more than they value success), so we cling to sunk costs and refuse good bets, and the standard prescription is to debias yourself back toward the spreadsheet. But the cowardly weighting is not a bug in the math. It is a correction to it. Expected value assumes you survive to keep averaging, and wherever a loss is absorbing (ruin, death, a torched reputation), [there is no long run left to average over](Decisions require ergodicity to succeed over time). The fear is survivors' math, written in by everyone who flinched and lived.

So the skill is not suppressing the fear, and it is not obeying it. It is checking which game you're in before deciding which voice gets the vote.

Quadrant
Trust the fear
Losses replay
tl

Money left on the table. The fear misfiring.

tr

Alive. The "bias" doing its actual job.

bl

Take every good bet. This is how casinos win.

br

The graveyard. Full of men whose math was right.

A loss ends the game
Trust the math

Source claim: Favorable odds justify a bet only when losing leaves you able to keep playing, so loss aversion is a built-in correction for games where it doesn't, not a bias to be trained away.

Related notes