🧠 Behavioral Economics: Loss Aversion, Framing and Prospect Theory

Pure Canvas 2D · real equations · in-page numerical checks
Numerical checks
The science

Prospect theory (Kahneman & Tversky) values outcomes as gains and losses from a reference point: v(x) = xα for gains, v(x) = −λ(−x)α for losses, with λ > 1 (loss aversion, about 2.25 in the classic estimate) and α ≈ 0.88. Probabilities are bent by w(p) = pγ / (pγ + (1−p)γ)1/γ (γ ≈ 0.61).

A 50/50 gamble +G / −L is accepted when Gα > λLα, i.e. G/L > λ1/α, so people typically want to win about 2.5× what they could lose. With λ log-normally distributed between people the accepting share is Φ((α ln(G/L) − ln λmed)/σ).

Framing: "200 saved" and "400 die" are the same outcome, but moving the reference point flips preferences from risk-averse (gains) to risk-seeking (losses). Choices there use a Gaussian noise term.

Simplifications: parameters are averages from lab studies and vary widely; people here don't learn, interact or trade, and real markets add many other effects. Educational model.