The Asymmetry
In 1979, Daniel Kahneman and Amos Tversky published "Prospect Theory: An Analysis of Decision under Risk." The paper described a simple finding with enormous consequences: the pain of losing $100 is approximately twice as intense as the pleasure of gaining $100. Not slightly more painful. Roughly double.
This is not a personality defect. It is a feature of human cognition so universal that it appears across cultures, age groups, and income levels. Kahneman and Tversky called it loss aversion, and it is one of the most replicated findings in the history of behavioural science. Losses and gains are not processed on the same scale. The value function is steeper for losses than for gains. Always.
Now place that asymmetry inside a casino — an environment specifically designed to create alternating sequences of gains and losses — and watch what happens.
Chasing Losses: The Predictable Response
You sit down at a slot machine with a $200 bankroll. After 30 minutes, you are down $80. Rationally, the next spin has exactly the same expected value as the first spin. Your past results have no influence on future outcomes. The maths does not change based on your session history.
But loss aversion does not operate on maths. It operates on feeling. That $80 loss registers with roughly twice the emotional weight that an $80 win would produce. The result is a powerful, visceral urge to continue playing — not because you expect to win, but because stopping now means crystallising the loss. As long as you are still playing, the loss is theoretical. The moment you stand up, it becomes real.
This is chasing losses, and it is not irrational in the way people think. It is a perfectly predictable response to loss aversion operating in a probabilistic environment. The player is not being stupid. The player is being human in a context that punishes human psychology.
Prof. Boston says
I ran an experiment with 200 undergraduates. Each received $20 in simulated chips and played a negative-EV game for 30 minutes. The students who were told their running total every five minutes quit 40% sooner than those who only saw it at the end. Visibility of losses accelerated rational quitting. The casinos know this too — which is why modern slot machines display wins prominently and bury net session losses in a sub-menu most players never open.
The Sunk Cost Trap
Loss aversion has a close cousin: the sunk cost fallacy. Money already lost in a session should have zero influence on your next decision — it is gone, irrecoverable, and irrelevant to the expected value of the next bet. But loss aversion makes sunk costs feel enormously relevant.
"I've already put $150 into this machine." That statement is economically meaningless. The machine does not know or care about your session total. But it feels meaningful because walking away from $150 in losses triggers twice the emotional weight of the $150 you might save by quitting. The sunk cost creates a psychological anchor that loss aversion pulls you toward.
This compounds over time. The deeper the loss, the stronger the pull to continue. The stronger the pull to continue, the deeper the loss. It is a feedback loop, and it runs on autopilot unless you have built a system to interrupt it.
The "I'm on a Roll" Problem
Loss aversion does not only affect losing sessions. It also distorts winning ones — in a less obvious but equally costly way. Suppose you are up $120. Prospect theory predicts that you will now become more risk-averse with your gains. You have something to lose, and losing it will hurt more than gaining another $120 would feel good.
But here is the twist: the casino environment reframes the situation. You are "on a roll." The variable reward schedule has been reinforcing your behaviour with wins. The near-miss effect has been maintaining engagement between wins. Your emotional state says keep going. Prospect theory says protect the gain. These impulses conflict — and in a high-arousal, fast-paced environment, the emotional state usually wins.
The result: players who refuse to walk away at a profit, give back their winnings, and then chase the losses they just created. The session starts with house money, transitions through loss aversion, and ends worse than it began. This is not rare. It is the default trajectory for unstructured sessions.
Stop-Loss, Stop-Win: The Architectural Solution
If loss aversion is an architecture problem, the solution must be architectural. Telling someone to "just walk away" is like telling someone to "just ignore" a variable ratio reinforcement schedule. It fundamentally misunderstands what they are up against.
This is why the bankroll calculator produces two outputs, not one. The stop-loss limit addresses chasing: you pre-commit to a maximum acceptable loss before your session begins, while loss aversion has no sunk costs to anchor onto. The stop-win target addresses the "on a roll" problem: you pre-commit to a profit level at which you leave, before the variable reward schedule starts whispering that more is coming.
Prof. Boston says
Professional poker players — people who make a living from gambling — almost universally use stop-loss rules. They will tell you the stop-win is actually harder to follow, because it fights a different instinct: greed amplified by confirmation bias. You are winning, so your brain tells you the strategy is working, that you have found a hot machine. There is no hot machine. There is a random number generator that happened to pay out, and a brain that evolved to see patterns where none exist. The stop-win is not about pessimism. It is about knowing which voice in your head to trust.
This approach is not theoretical. It is drawn directly from the behavioural economics literature on pre-commitment devices — strategies that constrain future choices to prevent predictable errors. Odysseus tying himself to the mast. Cutting up the credit card. Setting a stop-loss before the session starts. The principle is identical: your present self builds a constraint that your future self cannot easily override.
Seeing the Full System
Loss aversion does not operate alone. It stacks with near-miss engineering (each near-miss feels like a loss of something you almost had), variable ratio reinforcement (the schedule prevents you from finding a natural stopping point), and the sunk cost fallacy (past losses anchor your commitment to the session).
Understanding any one of these in isolation is useful. Understanding how they interact is the whole game. The casino does not rely on a single psychological lever. It deploys a system of interlocking biases, each one reinforcing the others. Your countermeasure needs to be equally systematic.
Start with the expected value framework — understand the maths. Layer in session management rules — address the behaviour. Then study the psychology that makes the behaviour hard to control. That is the order. Maths, systems, then awareness. Not the other way around.
Kahneman and Tversky did not discover loss aversion to make you feel bad about chasing losses. They discovered it so you could see the bias operating in real time and build structures to counteract it. The tools on this site exist because knowing about loss aversion is not enough. You need systems that work even when the bias is at full strength.