Prospect Theory: How We Really Feel Gains and Losses

Classical economics assumes you’re a cool calculating machine: you weigh every choice by your total wealth and pick the option with the highest expected payoff. It’s a lovely theory. It’s also wrong about almost everyone, almost always. In 1979 two psychologists — Daniel Kahneman and Amos Tversky — wrote down how humans actually decide under risk, and the paper (called Prospect Theory) went on to win a Nobel Prize and quietly explain half the dumb things investors do. This lesson is its core. Master these four ideas and you’ll catch yourself mid-mistake for the rest of your investing life.

The trailer — guess first

Prospect theory rests on four pillars. We’ll take them one at a time, each tied to something concrete you do with real money: (1) you judge outcomes from a reference point, not from your total wealth; (2) losses loom larger than gains; (3) you feel diminishing sensitivity, which makes you risk-seeking when you’re losing; and (4) you misweight probabilities, overrating the rare and underrating the likely. A fifth idea — framing — ties them together and shows how the wording of a choice flips your decision.

Pillar 1 — Reference points: gains and losses from where?

Stick one hand in ice water and the other in hot water, then plunge both into the same lukewarm bowl. The cold hand screams “hot!” and the warm hand screams “cold!” — same water, opposite sensations, because each hand judges from a different starting point. Your wallet works exactly the same way.

Reference point (definition): the baseline you mentally compare an outcome against. You don’t experience a portfolio value as a raw number on the scale of your lifetime net worth — you experience it as a gain or a loss relative to some anchor, usually the price you paid, the value at the start of the year, or the recent high. Classical theory says you should evaluate the final wealth level. Prospect theory says you actually evaluate the change from your reference point.

Worked example — the same million, two opposite feelings

Two people both end the day holding exactly 1,000,000 dollars.

Identical final wealth, opposite emotions. A classical economist would say they’re in the same situation and should feel the same. They emphatically do not — because each is measuring from a different reference point.

Why it matters for investing

Your most common reference point is your purchase price, and the market could not care less what you paid. A stock you bought at 100 that now trades at 70 isn’t “down 30%” in any sense the market respects — it’s simply worth 70, and the only question is whether 70 is cheap or expensive from here. But your brain has bolted a giant “−30%” sign to it, and that sign will tempt you to hold on “until it gets back to even.” That single distortion — refusing to evaluate a holding on its own merits because you’re anchored to what you paid — feeds the disposition effect you’ll meet in the next lesson.

Pillar 2 — Loss aversion: losses hurt about twice as much

Imagine finding a 50-dollar note on the pavement — nice little lift to your day. Now imagine instead losing a 50-dollar note out of your pocket — that one stings, and it’ll bug you for hours. Same 50 dollars, but the loss carries far more emotional weight than the gain. That asymmetry is loss aversion, and it’s the single most important idea in this lesson.

Loss aversion (definition): the pain of a loss is felt more intensely than the pleasure of an equal-sized gain. The widely cited rule of thumb is that losses feel about twice as strong — roughly 1.5× to 2.5×. Kahneman and Tversky’s 1992 estimate put the multiplier (called lambda, written λ) at about 2.25. Treat that as an illustrative average, not a law of physics: λ varies by person, by context, and by the size of the stake.

The value function below makes the asymmetry visible. It’s S-shaped, passing through your reference point at the origin: the curve for gains rises gently, but the curve for losses plunges steeply — same horizontal distance from the centre, much bigger vertical drop. Drag the slider and read off how much worse an equal loss feels.

Worked example — the value-function arithmetic

Prospect theory models the felt value of an outcome with a simple formula. For a gain of size $x$ and a loss of the same size, with curvature $\alpha \approx 0.88$ and loss-aversion $\lambda \approx 2.25$:

$v(\text{gain}) = x^{\alpha} \qquad v(\text{loss}) = -\lambda \cdot x^{\alpha}$

Take a stake of 100. The gain side gives a felt value of:

$v(+100) = 100^{0.88} \approx 57.5$

The loss side multiplies that same magnitude by λ:

$v(-100) = -2.25 \times 100^{0.88} \approx -129.4$

So a 100 gain feels like about +57.5 units of pleasure, while an identical 100 loss feels like about −129.4 units of pain. The ratio is 129.4 / 57.5 ≈ 2.25 — the loss stings roughly 2.25 times as much as the matching gain pleases. That’s the number that explains the coin-flip from the pretest: to make a fair coin worth flipping, the upside has to be about twice the downside, because the downside is weighted about twice as heavily.

Why it matters for investing

Loss aversion is the engine behind the most expensive investor reflexes:

• Panic-selling at the bottom. A 30% paper loss feels like a 60%-sized catastrophe, so people bail at exactly the wrong moment — converting a temporary, recoverable dip into a permanent, realised loss.
• Refusing to hold winners. A small gain feels good now; the fear of giving it back (a future loss from your new, higher reference point) tempts you to sell early and lock it in.
• Mispricing risk entirely. Because losing 20% hurts more than gaining 20% pleases, perfectly sensible long-horizon investors flee volatility they could easily have ridden out.

Pillar 3 — Diminishing sensitivity: why losing makes us gamble

Walk into a pitch-black room and light one candle — a dramatic difference, darkness to light. Now light a hundredth candle in an already bright room — you can barely tell. Each extra candle adds the same physical light, but you feel less and less of it. Money is the same: the jump from 10 to 20 dollars feels huge; the jump from 1,010 to 1,020 feels like nothing, even though both add exactly 10.

Diminishing sensitivity (definition): each additional unit of gain or loss is felt less than the one before. On the value-function curve this is the bending: the gain arm is concave (it flattens as gains grow) and the loss arm is convex (it flattens as losses grow). And here’s the consequence that matters most — that shape makes you risk-averse when you’re winning but risk-seeking when you’re losing.

Worked example — choosing to gamble to dodge a sure loss

You’re told you will definitely lose money, but you get to pick how:

The two options have the exact same expected value of −750. A purely rational, risk-neutral agent is indifferent. But faced with this, most people grab option B, the gamble — they’d rather risk an even bigger 1,000 loss for the slim hope of escaping clean, than accept the certain 750. Why? Because of diminishing sensitivity: on the convex loss arm, the felt difference between losing 750 and losing 1,000 is small, but the felt difference between “definitely losing” and “maybe losing nothing” is huge. The gamble’s tiny chance of a zero loss is worth more, emotionally, than the certain pain it spares you.

Now flip the signs to gains (a sure +750 versus a 75% shot at +1,000) and the preference reverses: most people take the sure 750, turning down the higher-expected-value-feeling gamble because, on the concave gain arm, a bird in the hand feels better than a slightly bigger bird in the bush. Risk-averse in gains, risk-seeking in losses — same person, same math, opposite behaviour, driven purely by which side of the reference point they’re on.

Why it matters for investing

This is the psychological fuel for doubling down on losers. A position is down 40% and staring at a “sure” realised loss if you sell. Diminishing sensitivity whispers that averaging down — buying more to lower your average cost — barely feels worse (losing a bit more on the convex arm is cheap in feeling-units) while dangling the seductive hope of “getting back to even.” So people pour good money after bad into deteriorating positions, the exact escalation behaviour you’ll dissect next lesson as the disposition effect. The rational move is to ignore the sunk cost and ask only “would I buy this today?” — but the curve is built to make that almost impossible to feel.

Pillar 4 — Probability weighting: the rare feels big, the likely feels small

Ask anyone why they bought a lottery ticket and they’ll say “well, someone has to win.” The actual chance is roughly one in hundreds of millions — practically zero — but it doesn’t feel like zero. It feels like a real, shimmering possibility. Meanwhile a 95%-likely outcome somehow never feels like the near-certainty it is. Humans don’t use raw probabilities to make decisions; we run them through a warped lens first.

Probability weighting (definition): instead of acting on a probability $p$, we act on a decision weight $w(p)$ — and that weight systematically overweights small probabilities and underweights moderate-to-large ones. The relationship is an inverse-S: it bows above the “fair” 45° line on the left (rare events) and below it across the middle and right (likely events).

The chart below plots the decision weight against the true probability. The straight diagonal is how a rational agent should weight each probability; the curved line is how people actually do. Drag the slider and watch the gap: way out at small probabilities the curve sits above the line (you give rare events more weight than they deserve), while across the meaty middle it sits below (you shrug off likely events).

Worked example — what overweighting a 1% chance costs

Suppose a lottery-like “story stock” has a genuine 1% chance of a 50× payoff and a 99% chance of going to zero. Its true expected value per 1 dollar staked is:

$(0.01 \times 50) + (0.99 \times 0) = 0.50$

Fifty cents of value for every dollar you put in — a terrible bet; you’d expect to lose half your money. But under probability weighting, that 1% chance might feel like, say, an 5–6% chance (the curve inflates it several-fold), while the 99% near-certainty of total loss gets quietly discounted. Emotionally the bet feels roughly break-even or even attractive, so people pile in. The math says “−50% expected return”; the warped weights say “lottery dream.” The weights win, and the wallet loses.

Why it matters for investing

Probability weighting explains two opposite-looking mistakes at once:

• Overpaying for the rare upside. Lottery stocks (penny stocks, meme longshots, far-out-of-the-money call options) are chronically overpriced relative to their true odds, precisely because the tiny jackpot probability gets inflated in everyone’s heads. You’re buying an overweighted dream.
• Overpaying for the rare downside. The same overweighting makes deep tail insurance (lottery-ticket “protection” against a market crash) systematically expensive — people pay too much to soothe a small, vividly imagined probability.
• Ignoring the boring likely. The flip side: high-probability, unglamorous outcomes — like “a low-cost index fund very probably beats your stock-picking over 20 years” — get underweighted, so people skip the near-sure thing for the exciting longshot.

Pillar 5 — Framing: same facts, different decision

Two yogurts. One says “90% fat-free,” the other “contains 10% fat.” Identical product, but the first one flies off the shelf. Nothing about reality changed — only the words — yet the words changed your choice. That’s framing, and it’s how the previous four pillars sneak into real decisions: by deciding, through wording, whether you’re looking at a gain or a loss.

Framing effect (definition): the way a choice is described — as a gain or as a loss — changes which option people prefer, even when the underlying outcomes are mathematically identical. Because loss aversion and diminishing sensitivity make us risk-averse about gains but risk-seeking about losses, simply re-labelling an outcome can flip your risk appetite.

Worked example — the Asian-disease problem

Kahneman and Tversky’s most famous demonstration. A disease threatens 600 people; you choose a programme.

Gain frame — described in terms of lives saved:

Faced with this wording, most people pick the sure A — risk-averse in the gain domain.

Loss frame — the identical outcomes, described in terms of deaths:

Now most people flip to the gamble D — risk-seeking in the loss domain. But look closely: A and C are the same outcome (200 live = 400 die), and B and D are the same gamble. Nothing changed except “saved” versus “die.” The framing alone reversed the majority preference.

Worked example — the same portfolio, two frames

You hold a fund. Two true statements about it:

• Gain frame: “It’s up 12% so far this year.” → You feel pleased, inclined to sit tight or add.
• Loss frame: “It’s still 8% below its 2021 peak.” → You feel disappointed, inclined to bail or avoid.

Same fund, same price, same day — but the first framing nudges you to buy and the second to sell. Brokers, fund marketers, and your own internal monologue all choose frames, and each frame quietly summons a different one of the pillars above (a loss frame wakes up loss aversion; a “since-the-bottom” frame resets your reference point). The defence is to re-frame every important choice at least twice — once as a gain, once as a loss — and check whether your decision survives both wordings. If it doesn’t, the frame is driving, not the facts.

The honest caveat

Prospect theory is one of the best-replicated findings in behavioural science — but it’s a description of how people choose, not a constant of nature or a licence to misbehave. Two cautions worth holding onto:

Putting it together

Five ideas, one picture: you judge from a reference point, feel losses about twice as hard, bend toward gambling when you’re already losing, misweight the odds, and let the framing of a choice decide which of those reflexes fires. Here’s the whole lesson at a glance:

A mixed recap pulling from every pillar:

Key Takeaways