Loss Aversion and Prospect Theory

Loss Aversion and Prospect Theory

Overview

People evaluate outcomes relative to a reference point (not absolute wealth), weight losses ~2.25x as heavily as equivalent gains, are risk-averse in gain frames and risk-seeking in loss frames, and distort probabilities (overweighting small, underweighting large). The same physical outcome feels different depending on framing — this skill diagnoses and corrects that asymmetry.

Composes with sunk-cost-fallacy, framing-effect, expected-value-and-kelly, anchoring, pricing-strategy.

When to Use

• A decision involves uncertainty and the chooser is visibly averse to a "loss" framing
• People are refusing positive-EV bets because the downside feels disproportionately bad
• Negotiations are stuck because concessions feel like losses from an anchored reference point
• A product launch, pricing, or incentive is producing unexpected adoption patterns
• Small-probability events are being over- or under-insured against
• An investor is holding a losing AI / Nvidia / semiconductor position waiting to "get back to breakeven," or is reacting to an AI-capex, AI-valuation, or AI-adoption drawdown (e.g. the DeepSeek shock) rather than re-deriving forward EV
• Someone says "loss aversion," "prospect theory," "reference point," "endowment effect," "status quo bias," "disposition effect"

Not when: the asymmetric weighting is rational (genuinely catastrophic stakes); the reference point is legitimate; the decision is small and one-shot.

Coaching Novices (Adaptive Front Door)

• Engine mode: user has a concrete case → run The Process directly.
• Coach mode: user is unfamiliar → guide step by step.

In Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.

1. One-line: before calling a risk choice irrational, identify the reference point and check if the decision flips when reframed gain vs. loss.
2. Check fit — if the loss is genuinely catastrophic and irreversible, asymmetric aversion is rational; use Kelly/antifragile, not debiasing.
3. Elicit the specific decision: what's being chosen, and what reference point makes one option feel like a "loss"?

[WAIT — do not advance until user responds]

1. Work through EV for each option; shift the reference point; test gain vs. loss reframing; flag over/underweighted probabilities.

[WAIT — do not advance until user responds]

1. Close: restate decision in EV terms and name explicitly how reference-point and probability-weighting influenced it.

[WAIT — do not advance until user responds]

The Process

Step 1 — Specify decision: options, probability × payoff distributions, reference point (explicit or implicit). Step 2 — Compute EV: Σ(probability × payoff) for each option; identify EV-dominant choice. Step 3 — Identify distortions: loss aversion (losses weighted >1x gains?), reference dependence (alternative reference points?), probability weighting (small overweighted? large underweighted?), diminishing sensitivity (large outcomes compressed?). Step 4 — Reframe and re-test: shift the reference point; restate as gain vs. loss; express probabilities numerically. If the decision flips, prospect-theory distortions are doing meaningful work. Step 5 — Choose decision rule: catastrophic+irreversible → respect loss aversion | moderate+repeatable → maximize EV | large+reversible → Kelly criterion | one-shot → add regret minimization. Step 6 — Document: chosen option, its EV, why it dominates, and which distortions were acknowledged/overridden.

Output Template

Decision: | Options (prob × payoff): | Reference point:
EV per option: | EV-dominant option:
Distortions: loss-aversion ratio | alternative reference points | probability weighting | diminishing sensitivity
Reframe test: decision under shifted reference point | gain vs. loss reframe
Stakes class + decision rule applied:
Final choice + acknowledged distortions + rationale:

→ Method in Action: Kahneman and Tversky's 1979 Prospect Theory · PGA Tour Par vs. Birdie Putts → 2026 lens: Holding AI Positions Through Drawdowns — the Disposition Effect (2023–2026)

Pack: Prospect Theory Patterns

→ Primary sources: references/sources.md

Common Rationalizations

[D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.

Red Flags

• Risk-aversion in gain frame + risk-seeking in loss frame for economically equivalent choices
• Reference point not made explicit; probability language verbal not numerical
• "Sure thing" chosen at significant EV cost
• Negotiation stuck at an arbitrarily-anchored reference point
• Investment held past rational exit because realizing a loss feels worse than its objective magnitude

Verification

• EV computed for each option
• Reference point made explicit; at least one alternative tested
• Decision re-tested under gain vs. loss reframing
• Probabilities stated numerically, not verbally
• Loss aversion respected (catastrophic) or overridden (moderate) deliberately, not by default
• If overriding, EV justification documented

---

Part of deciqAI Knowledge Skills — 227 open-source thinking skills that make rigor executable for AI agents. The same skills power every deciqAI agent, which runs them autonomously to operate your company. See it run → https://www.deciqai.com/c/loss-aversion-prospect-theory · ⭐ Star the repo → https://github.com/deciqAI/knowledge-skills · Contributions welcome.

Agents: latest version & machine-readable metadata → https://www.deciqai.com/s/loss-aversion-prospect-theory.json