Created on 2024-06-12

Thinking Fast and Slow

Rating

10/10 It is a fascinating book that dances on the edge of psychology and economics but applied to an everyday experience of an average human being who would like to be more aware of their opaque mind processes. Written by a true expert who has devoted his entire life to the topic of everyday human psychology, it is kind of a gift for the curious ones among us. The theory is intertwined with multiple dozens of research experiments that make the takeaways so much more clear and actionable. Although the general story is quite grim, we are apparently much more driven by emotional unconscious mind processes than we think, just being aware of this fact already gives us the ability to tap into our more rational side.

The question however is: do we want to be rational all the time? My girlfriend doesn’t think so, she sees beauty in being emotional, in following your gut, in enjoying the moment without thinking whether it’s optimal. And although there is a faint voice in my head telling me that ultimate rationality is what we should strive for, I must agree with her on this point. I think part of being human is being emotional and irrational - unlike agents in the Economic theory, and so not all the biases and mental shortcuts described in this book are as bad as they might seem. They are only bad if we rely on them in high-stakes situations (like deciding on the worldwide climate policy) and so the wisdom of this book mostly applies to such decisions.

The good news is, good news for my girlfriend that is, that even if I would like to be unboundedly rational, it is, for all intents and purposes, impossible. Engaging your rational side requires intense effort and will, it is exhausting, and unnatural in most situations (like intuitively catching a flying ball). So we will most likely forever use the mental shortcuts and heuristics described in this book, but the research presented here gives us the option to be aware of what’s happening in our heads, instead of just going with the flow, thinking we are the creators of our destiny.

Lastly, I think some people mind find the book repetitive. They key ideas are discussed at the very beginning and then circled back to at multiple occasions. But for a non-fictional book, I find this a good approach. It reinforces the ideas from multiple perspectives and gives you time and various contexts to think about their consequences. Plus repetition is good for retention.

Synopsis

400+ pages of mental shortcuts, biases, and heuristics we (mostly) unintentionally live by everyday - supported by stories, experiments, and results from volumes of research papers. It will open up your awareness about how you make decisions, how you form perceptions, and how you can call upon your rationality when its needed the most.

Notes

Intro

a book about judgments and choices
“We are often confident when we are wrong, and an objective observer is more likely to detect our errors than we are.”
doing research while walking and talking with a friend sounds like the pinnacle of intellectual human experience; modern day philosophers

Part 1

Chapter 1 - Systems 1 & 2

System 1 is the automatic, quick-witted part of your brain; the one you sometimes need to inhibit (like when a chocolate sits idly in your pantry) but also the one that sometimes saves your life (like when your bike skids on a frozen puddle but you instinctively save the fall)
System 2 is the deliberate thinker, the conscious attention allocator, and the complex situation navigator; it’s the part of you that helps you inhibit System 1 or think about what System 1 is intuitively telling you to do
2+2 is System 1 arithmetic, 1728 is *System’s 2 job
“System 1 sometimes answer an easier question than what it was asked”. E.g. asking “Should I buy Tesla stocks?” and you immediately responding yes, the response you gave is more likely the System’s 1 answer to the question “Do you like Tesla?” instead of the actual question “Should I buy Tesla stock?”. The latter requires some analysis and deliberation, the former can be answered on the spot. So System 1 can save your life, but it can also “make you blind to your own blindness.”
“He had an impression, but some of his impressions are illusions” - System’s 1 fault

Chapter 2 - Attention & Effort

“The most effortful forms of slow thinking are those that require you to think fast.”, like being asked a tough arithmetic question on the spot with time pressure to answer correctly
System 2 is like an electrical circuit at your home; it has limited capacity in how much mental load it can handle, if it gets overloaded, it selectively focuses on the most important task at hand and allocated limited residual capacity to the rest of the tasks. E.g. if you are driving and talking with a friend, a dangerous situation on the road will inhibit your ability to continue the conversation because all your attention is dedicated to handling your car through the situation
what’s a high effort activity for one person, can be low effort for another; e.g. chess game for me or for a chess master

Chapter 3 - Self-control and cognitive effort deplete the same mental resource

why potentially best thinking happens while walking - “a mild physical arousal of a walk may spill into greater mental alertness”
flow state = optimal human experience
“People who are cognitively busy are also more likely to make selfish choices, use sexist language and make superficial judgments in social situations” - hence “self control and cognitive effort are [both] forms of mental work” and System 2 has limited capacity to carry out both
“… when people believe a conclusion is true, they are also very likely to believe arguments that appear to support it, even when these arguments are unsound. If System 1 is involved, the conclusions come first and the arguments follow.” - terrible implications for everyday thinking, civil debate and societal discussion

Chapter 4 - Associative thinking & Ideomotor Effect

associative thinking happens whenever your brain registers some word or an idea and a whole network of associated concepts lights up in your brain. The problem is that most of these concepts are not registered by the conscious mind and so they go unnoticed. But this doesn’t mean you don’t act on them, you do, you are just not consciously aware that some association made you more likely to alter your later behavior or perception
ideomotor effect: the effect of an action being primed by an idea; e.g. unconsciously walking slower after being primed to think about elderly
we are Strangers to Ourselves, in a sense that we easily get primed to do certain actions without being consciously aware of the associative network of thinking that primed us to do it

Chapter 5 - Cognitive Ease & Strain

if your cognition is at ease, which is cause by having good mood, living a repeated experience, seeing information clearly in a big legible font, or being primed to an idea by associative thinking, then you are more likely to feel the information to be true and familiar, and you may yourself feel good and your thinking to be effortless
this is clearly a problem as it opens a loophole to trick you into accepting falsehoods by first putting you at ease and then feeding you information that you are more likely to accept at face value
moreover, “you have no simple way of tracing your feelings to their source”, hence you cannot tell if you are feeling at ease because something primed you to feel that way, or you are feeling at ease due to an unrelated reason to currently lived experience
“System 1 can respond to impressions of events of which System 2 is unaware”

Chapter 6 - Norms and Causality

we identify ourselves with System 2; when I say “I”, I mean System 2, my conscious self
even seemingly weird an improbable events can become perceived as completely normal due to how System 1 operates. Meeting a friend on the other side of the world while on a vacation would at first be perceived as completely abnormal, but if you meet the same friend by the same coincidence on some other vacation, System 1 will have you feel less surprised, as if it was okay for this friend to randomly pop up at the most improbable of places
“System 1, which understands language, has access to norms of categories, which specify the range of plausible values as well as the most typical cases.”
System 1 also creates stories and automatic causal connections that drive these stories. This is why we are so quick to mistake correlation for causation
“we see causality, just as directly as we see color” and “we are ready from birth to have impressions of causality, which do not depend on reasoning about patterns of causation” - In other words, System 1 sees causation, creates it out of thin air, without the need for System 2 to actually process whether causal effect took place

Chapter 7 - Jumping to Conclusions

“When uncertain, System 1 bets on an answer, and the bets are guided by experience.” - weighted with recency bias. Problem is, the bet is automatic and you are not aware of the alternatives that exist. They were considered unconsciously. “Uncertainty and doubt are the domain of System 2”
Confirmation bias: is the difference in your answers between “Is Sam friendly?” and “Is Sam unfriendly?”. The automatic first reaction to questions like these is to perform a so-called positive test strategy, where your System 2 tries to find supporting evidence for the question/claim at hand. It often finds some supporting evidence, and so you are more likely to answer affirmatively, conforming to the question/claim
so instead of our brains trying to find refutations to inputs we are receiving, they look for affirmations, which are easier to find
halo effect: we interpret an ambiguous or contradictory information such that it is coherent with the rest of the story/context. It works in a sequential manner, so the information we receive first sets the context for the halo to occur. There is a bias for favoring first impressions
the halo effect is often a problem in team discussions. So before an important issue is discussed at a meeting, it would be a good idea if all the participants first wrote down their opinion, and then discussed it, otherwise the people speaking later will be more likely to affirm what the earlier speakers said
System 1 wants coherent stories so it creates them instinctively and impulsively, not considering the amount or quality of data that support them. It follows a What You See Is All There Is (WYSIATI) heuristic which essentially makes it jump to conclusions and not seek alternative explanations or data

Chapter 8 - Prototypes & Sum-like variables

judgment heuristic: when we answer a question, e.g. “Should I invest in Tesla?”, by our judgment to a completely different but relevant alternative question, e.g. “Do you like Tesla?”
System 1 represents categories by a prototype, so it’s good at estimating averages but not sums or sum-like variables; you know what an approximate size of an elephant is but not how tall a stack of 10 elephants would be
System 1 is also able to match intensities across seemingly unrelated dimensions, e.g. you can say what would be the equivalent IQ score to someone being able to run 100m in 10 seconds (given you roughly now the distribution of scores on those two scales)

Chapter 9 - Substitutions & Endorsing of Emotions

when System 1 substitutes an easier question to the original one, it has to then use it’s ability to intensity match in order to answer the original question in its original units. E.g. if I ask you “How much would you contribute to save dolphins?” and your System 1 translates this to “How much do you like dolphins?”, before you can answer my question, you need to translate your “I like dolphins a lot” to a dollar value
you most often don’t notice when substitution of the question by System 1 happens; your awareness, System 2, is not present during this process
you also don’t notice that the original question might have been difficult because you automatically swapped it out for an easy question
your emotional attitude often drives your beliefs about risks and benefits of things; e.g. if you are emotionally positive about nuclear power, you are more likely to exaggerate its benefits and downplay its risks
part of the problem here is that your System 2 is more of an endorser than an enforcer, an apologists of System’s 1 feelings rather than a critic. How you feel about a thing is determined largely by System 1 and instead of System 2 challenging this by facts and analysis, it finds facts and arguments that fit within the emotional story generated by System 1
“an active, coherence-seeking System 1 suggests solutions to an undemanding System 2”

Part 2

Chapter 10 - Law of Small Numbers and Seeing Causation Everywhere

“intuitions about random sampling appear to satisfy the law of small numbers, which asserts that the law of large numbers applies to small numbers as well”
the following two statements are equivalent
- large samples are more precise than small samples
- small samples yield extreme results more often than large samples do
people are not sensitive to sample size, when they read a statement about the results of a poll, it doesn’t affect their beliefs about the result whether the sample is 150 or 1000. As long as it’s sufficiently big, the sample size is ignored, the story is what matters, not the source
we are thus prone to exaggerate consistency and coherence of what we see (WYSIATI); we essentially assume causation lot more frequently than we should
we pay more information to the content of the message than to the information about the reliability

Chapter 11 - Anchoring effect

anchoring effect occurs when people consider a particular value, some given number, in place of an unknown quantity they are asked to estimate
even if this anchor is completely random, it has a considerable effect on the final estimate of the people
e.g. asking “Was Ghandi older than 114 years when he died?” vs. “Was Ghandi older than 35 years when he died?” would most likely generate two significantly different averages that would correlate with the anchoring numbers used in these questions
anchoring happens for two reasons
- first, we use the anchor as an initial guess and then adjust from there until we are uncertain whether we should adjust further. This causes us to end our adjustments prematurely and stay in the vicinity of the anchor
- second, anchor has a priming effect, generating an associative web of related concepts that create a coherent story of why our final guess should be close to the given anchor. “System 2 works on data that is retrieved from memory, in an automatic and involuntary operation of System 1”
even random anchors can be as effective as informative anchors. So it’s not like we believe that anchors are necessary informative, we just cannot help ourselves to not be influenced by them
so since even completely random anchoring can have a large effect, “how free really are you???”
you should always assume that any number around you has had an anchoring effect, thus if stakes are high, you should mobilize your System 2 and analyze the situation thoroughly, being aware of the anchoring effect

Chapter 12 - Unexplained Unavailability Heuristic

availability heuristic occurs when we are asked about estimating a frequency of a certain category and our answer depends on the ease with which we are able to generate examples from that category
even worse, we often don’t really on the actual ease of generating examples, but our impression with which we should be able to generate examples. E.g. we intuitively know that generating words from letters TAPCERHOB will be easier than XUZONLCJMB without generating a single example from either set
so we estimated the size of a category as large not based on the actual number of examples retrieved but based on the ease with which we retrieve them
hence if we have a lot of trouble retrieving examples, but we expected to have little trouble, even if we retrieve a lot of examples eventually, we are likely to report small frequency for the category in question
so the availability heuristic is better called an unexplained unavailability heuristic

Chapter 13 - Emotion rather than reason

“How do I feel about it?” is the question we substitute and answer instead of the question “What do I think about it?”; hence we make choices and create opinions directly based off of our feelings, often without even knowing it
this is called an affect heuristic: making decision based on our feelings instead of objective assessment; “guided by emotion rather than by reason, easily swayed by trivial details” that just create a tidy coherent story in our minds

Chapter 14 - Stereotypes outweigh Base Rates

when we are given a detailed description of some specific exemplar and we are asked about what category this exemplar belongs to, we are much more likely to predict the category based on representativeness/stereotype rather than probability/base rates
e.g. if we have a description of a person that’s “shy and nerdy” and we are asked whether they are a computer science student or a humanities student, we will more likely answer the former while the latter is a much bigger group and thus statistically much more likely; we will do this even if we are told that the description has low reliability; we just cannot help ourselves, the System’s 1 lust for coherent stories is too strong
the only way to counter this is to employ your System 2, remind yourself of base rates, stay close to them and exercise self-monitoring and self-control

Chapter 15 - More detail makes scenarios less probable but more coherent

conjunction fallacy occurs when people judge a conjunction of two events as more likely than one of the events on its own when presented with a direct comparison
this again happens because of representativeness, stereotypes and System 1 creating a nice tidy (causal) story
e.g. if I tell you that Linda is an outspoken philosophy student concerned with issues of discrimination and then I ask you if it’s more likely that she’s a bank teller or bank teller that’s active in a feminist movement, you are more likely to go with the latter, while the former is by definition of logic more likely as it fully includes the latter category
this happens because we easily confuse coherence, plausibility and probability + we are more likely to answer according with out emotions that are inclined towards coherence and plausibility instead of probability
“adding detail to scenarios makes them more persuasive, but less likely to come true” - it plays into our thirst for coherence and our innate mechanism of representing categories by prototypes
laziness of System 2 is also part of the story, if we employed it, we would realize the logic should prevail but unless the stakes are high (our life or reputation depends on it), we are much more likely to just go with the intuitive, coherence-seeking System 1

Chapter 16 - Story-like statistical facts vs. plain statistics

we can trick ourselves into incorporating the base rates if they are presented as causal factors instead of statistical factors about the population; if they are more like stereotypes and less like facts
this is because System 1 likes causally linked stories, not statistical facts
this also tells us that if we want to learn some statistical fact, we are much more likely to understand it by seeing a few representative examples and then coming to the statistical conclusion by ourselves, instead of being given the statistical fact as is
this was shown to be true in the experiment that exposed the bystander effect

Chapter 17 - Regression to the mean

most causal interpretations of fluctuations in processes we observe are due to regression to the mean, not due to the causal story we tell ourselves
this is because every process we observe is part deterministic and part pure random luck, we don’t realize this and see processes as mostly deterministic, thus we attach causal stories to outcomes, when in fact they are very likely the result of the fluctuations in luck
regression to the mean is also equivalent to saying that there is imperfect correlation between the two successive outcomes we are observing; imperfect, because if it was perfect, then there would be no random/luck fluctuations
so if we are watching a high-jump competition and each athlete has two jumps, it is the most likely that the athletes with very bad first jumps will have better second jumps, and athletes with very good first jumps will have worse second jumps; this is the same as saying there is less than perfect correlation between the performance on the first and second jump. The relations between first and second jump is part deterministic (skill, training, height, temperature, etc.) and part luck; the luck causes most of the extreme events and so over time, its influence should cancel out, leaving us with the deterministic part (the mean)
the statements “highly intelligent women tend to marry men who are less intelligent than they are” and “the correlation between the intelligence scores of spouses is less than perfect” are equivalent statements, but the first one prompts us to come up with spurious causal explanations, while the second one is plain true fact that most people would find uninteresting; this get’s to show why we so often mistake causation for correlation

Chapter 18 - Intuitive predictions aren’t predictions at all; they are evaluation of present with no extrapolation

“some intuitions draw primarily on skill and expertise acquired by repeated experience” (chess-master-like intuition)
“other intuitions, which are sometimes subjectively indistinguishable from the first, arise from the operation of heuristics that often substitute an easy question for the harder one that was asked”
we should trust the first intuitions and make sure we aren’t fooled by the second
when we are asked about predicting something in the future we often substitute this for an easier question “what do I think now?”, hence most predictions are actually evaluations of current evidence
this substitution produces systematically biased predictions that fail to account for regression to the mean; the predictions are non-regressive, lacking the incorporation of uncertainty
here’s an example that tells us how to make better predictions. Imagine you are told that someone learned to read when they were 4 years old and now you are supposed to predict their university GPA. A prediction that is biased would be to just intensity match the fact that someone learned to read very early to GPA scale which is flawed as it doesn’t incorporate the uncertainty. A better way would be to:
- start with an estimate of an average GPA
- determine the GPA that matches your impression from the given evidence (the flawed prediction that’s actually a present evaluation)
- estimate the correlation between your evidence and the outcome, GPA in this case
- if the correlation is e.g. 0.3, produce your prediction by moving 30% of the distance between the base rate (average GPA) to the evidence-based GPA
- voila, this is your Bayesian-like prediction mechanism that should remove systemic bias
using this method you will still make errors, but they will be smaller as they won’t favor extremely high or low outcomes; it’s good for being right on average, not for predicting outliers; it’s not sexy
“the most valuable contribution of the corrective procedure […] is that it will require you to think about how much you [actually] know”
so in summary we should be wary of intuitive predictions because more likely than not they are biased, developed automatically by a coherence-seeking causality-hungry System 1 and much more extreme than what regression to the mean suggests

Part 3

Chapter 19 - Illusion of Understanding (hindsight/outcome bias)

narrative fallacy: we think we understand the world because System 1 is damn good storyteller. It constantly creates coherent, causal stories that are simple and concrete, assign a large role to talent, stupidity, and intentions than to luck, and focus on fee striking events that happened instead of the countless events that failed to happen (WYSIATI)
e.g. saying “Hitler loved dogs and children” is a crazy statement to hear because our System 1 prefers a coherent simple story that Hitler was an inside-out rotten person
“Our comforting conviction that the world makes sense rests on a secure foundation: our almost unlimited ability to ignore our ignorance.”
we are also unable to properly reconstruct past beliefs which makes us underestimate how surprised we were by past events. Once an event happens, we cannot reliably go back to before it happened and remember what we believed back then - this creates a hindsight/outcome bias
because of this illusion that we understand the past, we become overconfident in our ability to predict the future

Chapter 20 - Illusion of Validity/Confidence

illusion of validity: when we feel and act as if our predictions were valid because we are confident in our ability when in fact our predictions are at best a little better than random chance
this often occurs because confidence is a feeling that reflects the coherence of information and cognitive ease of processing - both System 1 attributes that easily fool us
“people are reluctant to infer specific from the general”
if people on Wall Street believe the efficient market (pricing) hypothesis, why do they trade? “Most of the buyers and sellers know that they have the same information; they exchange the stocks primarily because they have different opinions.”
the top experts are more disillusioned in their predictive accuracy than the well informed but not expert people. This is because the top experts have a very high illusion of their true skill which makes them overconfident in their predictions
so “high subject confidence is not to be trusted as an indicator of accuracy (low confidence could be more informative)”

Chapter 21 - Intuitions vs. Formulas

intuitions of experts operating in volatile environments (stock market, politics) are more likely than not to be less accurate than a simple equation/algorithm predicting the same outcome
this is because of illusion of validity and our inconsistency in making summary judgments of complex information; asked twice about the same judgment, even within minutes apart, will likely produce different conclusions - there are just too many unconscious factors influencing our judgments (e.g. little aches and pains, whether we are hungry, whether we are cold or warm etc.)
hence unreliable judgments cannot be valid predictors of anything
a way to escape this is to follow a standardized checklist for making predictions instead of expert intuition, employing System 2 instead of System 1
so to make a valid prediction, one can follow these steps instead of just relying on one’s judgment that’s volatile and easily influenced
- select few dimensions that are likely to be good predictors of what you are trying to forecast; preferably these dimensions are independent and you are fairly certain that you can assess them reliably and factually, e.g., by scoring them on a 5 point Likert scale
- for each dimension, create a list of scoring questions and a scoring system that will produce a final score for each dimension; have an idea of what answers would constitute a very weak or a very strong result for each question
- to avoid a halo effect, collect information on one dimension at a time and score it fully before moving onto the next one
- assess all the scores and decide solely on what the scores are telling you to do
if the above method is done by someone else, you can also ask them to “close their eyes and provide their intuition”, which will be another datapoint you can use for decision making
another quirk of human psyche is that we prefer mistakes to be made by other humans than to be made by algorithms. A patient dying because of algorithm’s fault produces more visceral feeling than the same death caused by the doctor’s wrong judgment

Chapter 22 - When Can We Trust Expert Intuition

expert intuition can be trusted if the expert is operating in a learnable, regular environment and he/she had enough time to learn the regularities; e.g. firefighters, chess players or surgeons can be trusted, but stock-pickers, real estate agents, and sports forecasters cannot
“intuition cannot be trusted in the absence of stable regularities in the environment”
intuition can be reduced to being just recall from memory:
- “The situation provided a cue; this cue has given the expert access to information stored in memory, and the information provides the answer. intuition is nothing more and nothing less than recognition.”
never take the degree of someone’s confidence in their predictions as a proxy for how much you can trust those predictions. The statements regarding one’s confidence are the result of System’s 1 coherent storytelling, hence they carry low external validity; don’t even trust your own confidence
anesthesiologists’ intuitions can be trusted as they have immediate feedback from their patients and the environment they operate it is very regular; radiologists on the other hand cannot be trusted as they often don’t receive feedback on their assessments, hence they have little opportunity to truly learn the regularities of the environment they operate in

Chapter 23 - Overly Optimistic Predictions

“a proper way to elicit information from a group is not by starting with a public discussion but by confidentially collecting each person’s judgment”
“people who have information about the individual case rarely feel the need to know the statistics of the class to which the case belongs” - probably true, unless you are a cancer patient and you want to know if the fight even makes sense…
when you are asked to predict some outcome about your specific situation you can adapt two distinct views to accomplish this prediction
- inside view: depends solely on the details of your situation and the progress you have made so far; is based on the current judgment of the situations extrapolated into the future without any regression to the mean; is most likely overly optimistic
- outside view: looks at the statistics of the reference category you find yourself in; assumes a certain baseline prediction of what you can expect on average for a case like yours; is non-specific to your case, but it’s the best guess in the absence of any details about your specific situation
the best way to predict is to combine these two views; use the baseline of the outside view and adjust it accordingly, up or down, to your specific situation
creating forecasts solely on the inside view commits the planning fallacy which creates overly optimistic forecasts that could benefit from taming by the baseline statistics
unfortunately, looking for the outside view is rarely natural and so it requires vigilance and effort to catch yourself before you make overly-optimistic predictions

Chapter 24 - Effects of Optimism on Capitalism and Success

on average “the financial benefits of self-employment are mediocre: given the same qualifications, people achieve higher average returns by selling their skills to employers than by setting out on their own”
entrepreneurial optimism is “widespread, stubborn and costly” - the chance of a small business surviving for five years in US is about 35%, yet small business entrepreneurs estimate their chances of success at 60%
when people are asked if they are above average on a certain activity, they are more likely to respond in the affirmative if they are merely moderately good at the activity (e.g., driving); if they perceive themselves as moderately bad, they are much more likely to say they are below average (e.g., starting conversations with strangers)
“an unbiased appreciation of uncertainty is the cornerstone of rationality”
although naive stubborn optimism often leads to demise of businesses, it’s likewise the key for the success of those who make it; the same holds true for academia. “Someone who lacks a delusional sense of significance will wilt in the face of repeated experiences of multiple small failures and rare successes, the fate of most researchers.”
a partial remedy for avoiding a failure due to overly optimistic predictions is doing an exercise of premortem: “Imagine that we are a year into the future. We implemented the plan as it stands now. The outcome was a disaster. Take 5 to 10 minutes to write a brief history of that disaster.” - this let’s you unleash your creativity in the direction you don’t let it go because you desperately want to succeed. The main virtue of this exercise is that it legitimizes doubts and let’s you adjust before you commit to an unrealistic plan

Part 4

Chapter 25 - Risk Aversion, Gambles and Marginal Utility

people don’t evaluate gambles based on expected values, instead they evaluate gambles based on psychological values of outcomes (their personal utilities)
in consequence people’s decisions are irrational because personal utilities weigh losses and uncertainty much more heavily than gains and certainty; people are also risk averse because they have decreasing marginal utility from wealth, each additional increase in wealth has a smaller perceived value thus they are not willing to risk much for small perceived gains
risk aversion flips to risk seeking only if all your options are bad; if you are given two bad options, you are more likely to pick the riskier one since there’s a high chance you’ll loose something anyway, so might as well risk it
theory induced blindness: “once you have accepted a theory and used it as a tool in your thinking, it is extraordinarily difficult to notice its flaws”

Chapter 26 - Prospect Theory: Neutral reference points, risk aversion and diminishing sensitivity

in gambles, and decisions in general, the choices are made in comparison to the reference point from which the options are evaluated. However, you don’t make decision based on the evaluation of your reference point (wealth), it just plays a factor in how much risk you are willing to respect and for what amount
in general people hate loosing and love winning, and they hate loosing much more than they love winning. The ratio between how much you need to gain to offset a negative feeling experienced by loss of a single unit is estimated between 1.5 and 2.5 the amount
the principle of diminishing marginal utility or diminishing sensitivity is just that the subjective difference between 100 euros and 200 euros is much larger than 900 and 1,000 euros; we essentially loosely think in percentages or ratios
when we turn down a risky gamble, the decision is made by System 2, but the inputs for the decision making are the automatically triggered fear and risk aversion of System 1

Chapter 27 - Status Quo and Indifference Curves

in elementary economics there is a theoretical concept of an indifference curve. It depicts a curve that shows a tradeoff between two commodities where an economic agent is indifferent to where it stands on this curve as he extracts equal utility from any of the combinations on this curve
in reality we are not indifferent and when we are offered a move alongside this curve we are much more likely to turn it down because a move always implies some loss. Because we weigh losses more heavily than gains, we are more prone to prefer the status quo, current situation, than a change
a response to a loss is stronger than a response to a corresponding game (1.5-2.5 times stronger)
the loss aversion however only applies to goods and money we intend to keep “for use” instead “for trade”. So there is no loss aversion in basic trade exchanges, like buying shoes or groceries
loss aversion is part of System 1

Chapter 28 - Goals are also Reference Points

“we are driven more strongly to avoid losses than to achieve gains”
a reference point is usually our current state but it can also be a future goal: exceeding this goal is a gain, not achieving it is perceived as a loss. This is why we set goals and why we are more likely to just meet our goals rather than exceed them. Once we cover the fearful possibility of a loss, we are contempt and the prospective gain of exceeding our goal is not as desirable as avoiding the loss of not meeting the goal
there is also a role of fairness in economic transactions, it’s not just all purely driven by supply and demand. Studies show that people find it unacceptable when firms use their market power to impose losses on others

Chapter 29 - Overweighting and Underweighting of Probabilities

most of the time “you are just an observer to global evaluation that your System 1 delivers”
a change of a chance from 0% to 5% has a much bigger emotional effect than 5% to 10%. Same is true for 95% to 100% versus 60% to 65%; this is due to possibility effect and certainty effect; the more emotionally salient changes register as a qualitative shift, while their counterparts are mere quantitative shifts
the possibility effect results in humans weighing unlikely outcomes more than they deserve (overweighting), while the certainty effect results in underweighting
“the psychological effect between 95% chance of disaster and certainty of disaster is HUGE”
in an experiment researchers tried to estimate the psychological weights people assign to different probabilities, they found that 0 and 100 percent probable events have correct weights 0 and 100, but in between, there is overweighting of low probabilities and underweighting of high probabilities, e.g. 1% -> 5.5, 2% -> 8.1 and 95% -> 79.3, 99% -> 91.2
the weights assigned to the range 5% to 95% was only 13.2 to 79.3, so about 2/3 of what the actual spread is
this mismatch between actual probabilities and implicit weights creates a 4 quadrant (fourfold) pattern of preferences:

	GAINS	LOSSES
HIGH PROBABILITY	95% chance to win $10,000	95% chance to lose $10,000
	Fear of disappointment	Hope to avoid loss
	RISK AVERSE	RISK SEEKING
	Accept unfavorable settlement	Reject favorable settlement
LOW PROBABILITY	5% chance to win $10,000	5% chance to lose $10,000
	Hope of large gain	Fear of large loss
	RISK SEEKING	RISK AVERSE
	Reject favorable settlement	Accept unfavorable settlement

the bottom left quadrant is where people buy lottery tickets; not because they believe they can win but because by buying they get the right to pleasantly dream about winning
the top right corner is where biggest misfortunes happen, it’s the spot where a company should decide to cut its loses and accept defeat but instead it pours all its remaining money into some new gimmick with a sliver of hope for making it out of the ditch
in each scenario we are essentially paying a premium for our overweighting or underweighting of probabilities - this gets costly in the long run!

Chapter 30 - Rare Events + Frequencies vs. Probabilities

with rare events the actual probability is inconsequential, for our fantasies to get going, only the possibility matters
people both overestimate probabilities of unlikely events and overweight them in their decisions; both of these happen due to the same string of mental events: focused attention + confirmation bias (WYSIATI) + cognitive ease
the probabilities of rare events are most likely to be estimated when the alternative is not fully specified; e.g. when asked about a probability of a team winning an NBA or a team from east coast winning an NBA, the first is more likely to be overestimated as the alternative is not clear, while in the latter case, there’s a clear alternative of a west coast team winning the tournament
so when the alternative is unclear, diffuse, we are more likely to overestimate the probabilities of success; this feeds back into the planning fallacy where we are more likely to predict our success precisely because we cannot fully imagine all the ways things could go wrong
denominator neglect: the bias to favor options with higher numerator even if the entire fraction has lower probability, e.g. favor winning something with chances 8/100 versus 1/10 (balls in the urn)
the denominator neglect works mostly when we state small probabilities in terms of frequencies (how many units out of 1,000) than in more abstract chances, risks or probabilities (“how likely” formulation)
a chance of 0.001% of fatal side effects from a vaccine is generally underweighted compared to same risk being presented as 1 in a 100,000 people experience fatal side effects; the 999,999 people somehow fade into the background and the focus is shifted on the one salient person
the frequency formulations are somehow more “vivid” and in turn they produce higher weighting factors for the same negligible probabilities; this is because vivid descriptions generate cognitive ease and richer more salient associative networks
the overweighing of low events usually happens when one is given a description to read and decide upon it. The description in general, vivid or not, influences our System 1 and since it only sees what there is, it is more likely to overweight the event. When the weighing is supposed to be generated based on experience, it’s generally more closer to reality and often underweighted with rare events. This is because rare events are rarely experienced, hence their chances are underweighted - that’s why we are not worrying about climate change as a society.

Chapter 31 - Broad View (Risk Policies)

“we have neither the inclination nor the mental resources to enforce consistency on our preferences, and our preferences are not magically set to be coherent”
if we face a set of decision events we can either adapt narrow framing and consider them in sequential order (as independent) or broad framing and consider all the decisions at once (still as independent but as a set, kind of a long-term view)
what usually happens is that we consider independent decision events one at a time and although we find them unfavorable on their own, in a long-run of such events we would register a gain - the “you win some, you loose some” heuristic, if we thus accepted small risks and fought our risk aversion, we could benefit in the long run - this applies to small financial bets that are inconsequential to our wealth if we loose them, like stock market investing with money that’s put aside for that purpose
the trick to counter risk aversion in potentially financially positive long run bets is to “think like a trader”; you should adopt a risk policy: a broad view that embeds a particular risky choice in a set of similar choices

Chapter 32 - Regret and Blame (Deviations from Norm)

“money is a proxy on a scale of self-regard and achievement”
“we are biased against actions that could lead to regret” and “we refuse to cut losses when doing so would admit failure”
there is a big difference in our perception of “commission vs. omission”, we place much more value on the outcomes that happen due to what we did than due to what we did NOT do; the sense of responsibility for our actions is greater than our in-actions
when it comes to selling stocks, people prefer to sell “winners”, stocks that gained in value, rather than “losers”, stocks that lost value. This is even though financially it would be beneficial to sell losers as there are no (or even negative) taxes on these sales
“regret and blame are both evoked by comparison to a norm” - if you do something out of norm and it leads to a bad outcome you are more likely to experience regret than if a normal action would lead to the same bad outcome
“consumers who are reminded that they may feel regret as a result of their choices show an increased preference for conventional options, favoring brand names over generics”
doctors have an emotional preference for sticking to the standard treatment option as deviating from a norm poses a higher risk of feeling regret, blame and potentially litigations
Kahneman’s policy to minimize regret is the following: “be either very thorough or completely casual when making a decision with long-term consequences. Hindsight is worse when you think a little, just enough to tell yourself later: I almost made a better choice”

Chapter 33 - Single vs. Joint Evaluations: Decision Reversals

regret, blame and also poignancy (a sorrow or pathos) are counterfactual feelings; evoked by the though “only if I did/did not do XYZ”
sometimes we reverse our decisions depending on whether we create them in joint evaluation of choices or in single evaluations E.g. if you are asked about what amount would you contribute to saving dolphins, you will come up with an amount that emotionally matches your degree of caring about dolphins and past amounts you contributed to similar projects. Now if you are asked to contribute to saving farmers from skin cancer, you are most likely going to come up with a lower amount, as not many people particularly like farmers or donate to similar causes. However, if you were asked to donate to both saving farmers and dolphins, you are more likely to contribute higher amount to farmers than dolphins, because farmers are humans and hence more valuable. This is an example of decision reversal. In single evaluation of options you valued dolphins more than farmers, but in joint evaluation the values reversed
the single evaluations are more likely to be determined by an emotional evaluation of System 1 while the comparison in joint evaluations is carried out by more rational System 2
the problem is that world runs on a between-subjects experiments rather than within-subjects. You usually only see a single options and it’s hard to be diligent with your System 2 and generate counterfactuals yourself. Thus you are often swayed by the option you see, committing to an emotional decision created by System 1 (WYSIATI)
“judgments and preferences are coherent within categories but potentially incoherent when the objects evaluated belong to different categories” -> decision about the amount donated to dolphins vs. hedgehogs will be coherent when made in wither single or joint evaluation, but dolphins vs. farmers might not be
“you should be vary of joint evaluation when someone who controls what you see has a vested interest in what you choose” (marketing) however, except for this case, your decisions are believed to be more rational and stable in joint rather than in single evaluations (System 2 vs. System 1)

Chapter 34 - Frame-bound vs. Reality-bound Decisions

framing matters!
“a bad outcome is much more acceptable if it is framed as a cost of a lottery ticket that did not win than if it is simply described as a losing gamble” -> losses evoke stronger negative feelings than costs
COSTS ARE NOT LOSSES (a mantra from Richard Thaler to keep in mind)
people will more readily forgo a discount than pay a surcharge; although economically equivalent, the two are not emotionally equivalent
when doctors where asked about recommending surgery that A: has a one-month survival of 90% or B: has a 10% mortality in the first month, they were for surgery 84% of the time in case A but only 50% of the time in case B, although the two are equivalent!!!
if the doctors where reality-bound decision makers (like in Economic theory), they would make the same decision regardless of the framing, this is clearly not the case
hence our preferences are often frame-bound instead of reality-bound. Of course not everyone is susceptible to framing, the individuals who aren’t are usually more accustomed to employing their System 2
the act of framing shouldn’t be seen as an intervention that masks or distorts the underlying preference; our preferences are about the framed problem, our moral intuitions are about the descriptions, not the substance!
thus broader frames and more inclusive accounts of the described situations usually lead to more rational decisions - these frames are closer to reality

Part 5

Chapter 35 - Discrepancy between Experiences and Memories

“rational agents are expected to know their tastes, both present and future, and they are supposed to make good decisions that will maximize their interests”
inside ourselves we have two distinct selves: the experiencing self and the remembering self; in general we wish to make decisions that maximize the utility of the experiencing selves, however we make these based on the memories remembered by the remembering self, if this self is biased in how it create memories, we we constantly sell ourselves short in our decision making - and the remembering self is indeed biased
our memories are biased in two ways:
- peak-end average rule: we forge our memories by the average of the intensity at the peak of the experience and the intensity at the end of it
- duration neglect: we ignore the length of the experience, only the peak and the end matters
“what we learn from the past is to maximize the qualities of our future memories, not necessarily of our future experience. This is the tyranny of the remembering self.” - but is this so bad? we often do challenging activities not for the experience but for the memories and the memories at the end make us who we are

Chapter 36 - Life is a Story of Peaks with Duration Neglect

“a story is about significant events and memorable moments, not about time passing”
“caring for people often takes the form of concern for the quality of their stories, not for their feelings… we can be deeply moved even by events that change the stories of people who are already dead”
hence stories are about peaks (and ends) and a good story is not judged by the number of good experiences (a sum variable) but rather by the average of the experiences lived
people choose by memory when they decide whether or not to repeat an experience -> which is biased by both the duration neglect and peak-end average rule
“I am my remembering self, and the experiencing self, who does my living, is like a stranger to me”

Chapter 37 - Determinants of Experienced Well-Being

definition of happiness according to Kahneman:
- you spent most of your time engaged in activities that you would rather continue than stop, little time in situations you wished to escape, and - very important because life is short - not too much time in a neutral state in which you would not care either way
sign of having a good time (being in flow): “resistance to interruptions”
how to improve experienced well-being (instead of remembered well-being):
- spent more time in active leisure (socializing, sports) instead of passive leisure (TV watching)
- don’t combine different activities, e.g. eating with TV watching, as combining activities dilutes pleasure of either of them
- spend more time with friends/relatives; people you care about and who care about you (best predictor of the feelings of your day)
- don’t be poor but also don’t be incredibly rich; being poor is a source of stress, being rich diminishes one ability to enjoy the small pleasures of life - the sweet spot was $75,000 in US (in 2010)
measures of life satisfaction do not measure experienced well-being, they measure the memory of well-being, which is flawed towards the peak emotions and neglects duration

Chapter 38 - “Miswanting” and the Focusing Illusion

responses to global well-being questions should be taken with a grain of salt; they are often answered by substitution and using the mood heuristic; answering how your mood is right now, instead of how your objective well-being has been for the past few weeks/months/years
the score that you quickly assign to your life’s satisfaction is not a carefully weighted combination of all the domains of your life, instead it is a quick evaluation of the most influential and readily available events
“a disposition for well-being is as heritable as height or intelligence”
your well-being is often not determined by what you objectively achieves but by whether what you achieved (or did not achieved) was your goal all along. When comparing people’s satisfaction as a product of financial wealth, those who set out to be wealthy and achieved it reported higher levels of life satisfaction than those who were equally wealthy but it wasn’t their goal all along; same but in the opposite directions for the poor
focusing illusions: nothing in life is as important as you think it is when you are thinking about it
when asked whether you like your car, you are more likely to exaggerate your feelings about it, due to the focusing illusion, when in fact most days it’s just a mean to and end and you feel quite neutral about what vehicle you drive
“adaptation to a new situation, whether good or bad, consists in large part of thinking less and less about it” So most long-term circumstances of life (marriage, chronic illness) are in fact part-time states of mind
experience sampling studies show no significant differences in experienced well-being between fully able and disabled people, yet, according to their remembering self, disabled people would be willing to pay large amounts of money and trade years of their life for living fully abled life - the consequence of peak-end rule, duration neglect and focusing illusions
this is a type of miswanting which occurs when we want something according to our remembering self but that desire doesn’t reflect the true state of our experiencing self
the focusing illusion results in favoring goods and experiences that are initially highly exciting but fade out with time, it also causes the mildly exciting experiences that last a long time to be underappreciated; it’s a mistake of focusing on the selected moments, transitional moments, instead of on the long-run affect

Conclusions

Experiencing vs. Remembering Self
- “the central fact of our existence is that time is the ultimate finite resource, but the remembering self ignores that reality. The neglect of duration combined with the peak-end rule causes a bias that favors short period of intense joy over a long period of moderate happiness.”
- aiming to make rational decisions based on the experiencing self is a noble goal but people care about their remembering self, and thus about their story, which does not square with optimizing for weighted sum of experiences
- for true happiness, both the experiencing and the remembering self must be considered, since their interest do not always coincide
Econs vs. Humans
- Econs are internally consistent, Humans are not
- Econs are rational (logically coherent), Humans are not
- thus humans require nudges to make more accurate judgments. These nudges can be provided by the state/the system but there is a slippery slope between adequate nudges and state-serving control of the populous
- but things like opt-out programs from organ donations, instead of opt-in; food labels that are framed such that they nudge people towards healthier choices (10% fat vs. 90% fat-free); and gas millage units that promote better vehicle choices (miles per gallon vs. gallons per mile) are all great unintrusive ways to get Humans closer to Econs
System 1 vs. System 2
- “System 2 is not a paragon of rationality” - “it’s abilities are limited and so is the knowledge to which it has access” - often it acts upon the prompts provided by System 1
- System 1 is the origin of most of the things we do wrong (mental shortcuts, biases, heuristics) but also of much that we do right (quick reactions, gut feelings), and most of the time, we do the things right rather than wrong
- learning about the mistakes of System 1 only allows us to spot them more often and correct for them, not avoid them entirely; unfortunately it’s also much easier to spot other people’s mental mishaps than our own