Sunday, March 28, 2010

My Models of Thinking, v. 1.3

I've been thinking about some speeches Charlie Munger gave in which he outlines one source of his excellence as being from the ability to apply all sorts of mental models. He notes that "man with a hammer syndrome" - to a man with a hammer, the whole world looks like a nail - can make people torture reality into the shape that fits their framework of thinking, instead of the other way around. He notes that cultivating a variety of mental models - different structures to use when attacking a problem - is extremely useful, and even more so in list form, so that you can look at the list and apply the relevant ones.
I am going to cultivate a list of models here. It's not perfect, but it's a reference list for how to think of solving problems. They're roughly categorized but in no particular order.

IF YOU CAN THINK OF GOOD ONES FOR ME TO LEARN ABOUT OR ADD, PLEASE LET ME KNOW!!!! I'm always looking to learn new things.

1. Arithmetic/Algebra. Addition, subtraction, multiplication, division, exponents and powers, logarithms.
2. Arithmetic with orders of magnitude - addition, subtraction, division, multiplication, powers.
3. Matrix Algebra
4. Expected value
5. Combinatorics - Mental math of combinations and permutations with and without replacement
6. Conditional and Bayesian statistics
7. Basic calculus (Integration, differentiation, limits, sequences and series)
8. Differential equations and multiple equilibria
9. Decision trees and probability-weighted outcomes
10. Distributions: Mean, Standard Deviation, Skew, Kurtosis (fat tails), median/trimmed mean, IQR, etc.
11. Compounding, and how small effects become huge over time.
12. Law of Large Numbers
13. Inversion... if it's not clear going forwards, start with the result you seek and get to the start.
14. Statistical sampling
15. Data-mining as a method of model construction as opposed to model construction and testing - using inputs and results to build a model instead of inputs and a model to see if they match results
16. The more leeway or bias in the construction of a study, the more likely it is to reflect funding incentives or a prevailing bias and the less likely it is to be true.
17. Sources of bias: survivorship bias, omitted variable bias, self-selection bias, other sampling bias, recency bias, endogeneity bias, etc.

1. Costly current investment for greater future gain
2. Market as an ecosystem of independent moving parts that affect each other
3. Diminishing returns to scale
4. Zero-sum games
5. Nash equilibria, the prisoner's dilemma and game theory
6. Economies and diseconomies of scale, and when they happen.
7. Specialization - Adam Smith's pinmaker, or Henry Ford's factory.
8. Capital intensity and scalability
9. The tiptoe advantage. A new technology is, for a business, like standing on tiptoe at a football game. It works great when you're the only one doing it, but sooner or later everyone is standing on tiptoe, and you've lost the advantage. If it was costly to stand on tiptoe, you shouldn't have done it (think Buffett's closing of the Berkshire Hathaway mills)
10. INCENTIVES. Possibly the most powerful on this entire list.
11. Time value of money.
12. Tragedy of the Commons
13. Ricardian equivalence and comparative advantage- trade can make everyone better off.
14. Rational people think at the margin
15. All decisions require tradeoffs
16. Opportunity costs and actual costs - the cost of something is what you give up to get it.
17. Fixed vs variable costs.
18. The use of markets to organize intellectual or economic activity.
19. Supply and Demand, and elasticity of supply and demand.
20. Income and substitution effects.
21. Externalities - positive, negative and positional.
22. Monopoly and monopsony, and oligopoly and oligopsony
23. Local monopoly and transport costs.
24. Second- and third-order effects
25. The process of theoretical modeling and then regression testing to verify
26. Barriers to entry - patents, brands, physical resources, scale, entrenched distribution, network effects, predatory undercutting pricing ability from financial resources.
27. Hyperbolic discounting and discipline.
28. Mutually assured destruction as a deterrent.

1. Diversification as a measure of reducing risk
2. Familiarity with operations as a measure of reducing risk (often counter to diversification).
3. Volatility vs value as a measure of assessing risk
4. Accounting as a language of business, and the fact that its an attempt to standardize representation of reality instead of being reality. 
5. The 80/20 rule.
6. Time-horizon arbitrage
7. Tax-avoidance.
8. Schumpeter's competitive/creative destruction, and how either you do it or someone else does.
9. Circle of competence - do what you know how to do well, and don't stray. If this means taking a lot of pitches, take them, don't swing.
10. Elevator pitch - if you can't boil a complex situation down to three key points, you don't understand it (contrary to irreducible complexity in physics)
11. Premortems and postmortems ("If something is going to go wrong, what will it be?" or "What went wrong?")
12. Instead of going for a 7 foot hurdle, find 7 one foot hurdles. Doing something really hard or doing things perfectly shouldn't always be necessary.
13. Parimutuel betting, and the importance of the ante.
14. Mean reversion
15. Secular vs cyclical trends.
16. Whole is greater than/less than the sum of its parts.
17. You need to put in 10,000 hours of practice before it's time for you to be judged if you want to excel. Gladwell.
18. Self-fulfilling prophecy (like autocatalysis in chemistry)- if your price goes up irrationally, you can raise more capital and justify it.
19. People buy fluent products, where the decision-making is easy - easy to use, features described clearly (and in a clear font). Ease of use makes things popular, because ease of use gives us pleasure
20. Better to initiate a price war than respond to one unless you're already a hard discounter by reputation- Initiating a price war makes customers more price sensitive, but the initiator gains a better price image while responders don't. (Still better not to have one)

1. Flowcharting for decisionmaking
2. Style vs substance for decisionmaking
3. Quality vs Quantity
4. Reciprocity tendency
5. Social proof, and the danger of groupthink
6. Commitment and consistency
7. Liking and familiarity as a compliance tactic
8. The power of authority, or its appearance
9. Scarcity and reactions to scarcity
10. Loss aversion
11. Irreducible complexity. People oversimplify because it's comfortable, but there are often unavoidably lots of things out of your control
12. Over- or under-emphasis on extreme values. Goes for both probability scenarios (in which case it's a black swan) or also choosing a level of service to buy or provide (we tend towards getting either "the nicest" or "the cheapest").
13. Paralysis by analysis. Stick to fundamentals ("First-best" vs "second-best" economists, for example).
14. The importance of communicating "why" in addition to "who, what, where, when and how"
15. Pavlovian conditioning, and the importance of imagery, consistency and caution about association. Classic conditioning (stimulus elicits emotional response - coke ads), Instrumental conditioning (positive or negative reinforcement of a behavior -  why increasing prices can increase sales because its a quality measure), operant conditioning (associative learning and teaching).
16. Denial
17. Lollapalooza effect - when multiple models combine in the same direction to push themselves farther than they could individually
18. The power of questioning: people are more likely to believe an answer they give than they are to believe one that you give
19. Cognitive dissonance and taking shortcuts to resolve it
20. emotional persistence - if you do something in an emotional state in a certain way, you're more likely to do it in that way again, long after the emotional state has gone
21. don't make emotional decisions
22. Sleeping on decisions helps your brain work it out while you're asleep in a less emotional way
23. Trust your gut after ruminating, but not your first instinct.
24. Everything takes longer than you think it does, to get an accurate estimate, ask someone else who has done it. Your situation is not different.
25. Anchoring and perceptual contrast.
26. People always perceive bias against them, irrelevant of whether there's actually bias against them.
27. Conformity plateaus in groups of 3-5 people large - after that, additional people don't increase conformity. A single consistent dissenter can crush conformity, but they must be consistent.
28. People conform more in a good mood. Compliance technicians use "danger-then-relief" mechanisms to increase conformity (make everyone feel endangered and then give them relief so they don't question).
29. Different individuals have different levels of conformity and preference for structure.
30. Foreigners or members of an "other" group are difficult to think about and assess.
31. Fluency and ease of use/thought makes us make faster, more emotional and less thoughtful decisions.
32. However, thinking abstractly instead of concretely (or even just being in an abstract state of mind, instead of a concrete one) makes us exercise more self-control, because we're not thinking of the tempting option right in front of us. Self-affirmation also increases self-control, because we think of ourselves as strong in our values. (Glucose also increases self-control).
33. Synthesizing fluency with scarcity... hard to get works best if it looks like it's hard to get for everyone ELSE, but for right now this second, it's not hard to get for you.
34. Surprise is a compliance tactic, because people don't have a chance to mount an emotional or intellectual defense or rationalization of their own stance.
35. People come to believe their own lies if they are repeated often enough.
36. Choice is demotivating. The more choices you have, the less you want to choose. The more alternate options you have on something, the less you work on the option you select.
37. Psychological Priming
38. Mentally separating from something - psychological distance in either a geographic or chronological sense - may make you better at insight problems.
39. Absurdist stimulation or forceful combination of opposites can spark creativity, so go off topic and ponder the impossible. Similarly, the path of most resistance sparks creativity, while least resistance stifles it.
40. Reconceptualizing problems in different forms or structures can lead to very different solutions.

1. Critical Mass - you need a certain amount of something to get it moving, but once gathered together, it explodes.
2. Break points - in any complex system, there are going to be certain points which are more vulnerable than others, and the system is as strong as its break points
3. Foundational construction - you can only build as high as the foundation will support.
4. Positive and negative feedback loops - self-correcting systems and self-destroying systems.
5. Backup systems
6. Theory of equilibria, and static vs dynamic equilibria.

Computer Science
 1. Orthogonal operations - an instruction set where each command does one thing, and no other command does what any other one does. (as opposed to multiple ways to do something)
2. Refresh rate - the rate of updated information affects what you can do with that information

1. Evolution and survival of the fittest (natural selection)
2. Vertical and horizontal gene transfer
3. Catalysis.
4. Autocatalysis - something produces its own catalyst.
5. Survival by differentiation and niche filling.
6. The Scientific Method and Experimentation

1. Differential diagnosis

1. "You can only get rich once."
2. In a repeated game, those who truly exert effort and discipline will keep going until they find a way to succeed. 
3. The idea of giving people a second chance
4. The idea of prolonging the inevitable
5. People don't change, but the way they manifest who they are changes and matures.
6. People who consider themselves moral, or perceive what they do is moral, are more likely to be immoral, because complacency means they don't seek to be better. Alternately, if people feel like they "deserve" reward for doing something right, they'll go do something wrong to reward themselves. Backed by studies on "halo of green consumerism".
7. Validity vs Soundness

Sociology, Anthropology and Culture
1. The idea of a social network - if your friends are fat, you become fat. If your friends have fat friends, you're more likely to be fat also (first and also SECOND order associations). See Christakis.
2. Some cultures are inherently more conformist (typically East Asian) and some are more individualistic. (Verified in experiment, Kim & Markus 1999, Smith & Bond 1993)
3. Overconfidence serves as a signal to others about one's own skill - either a bias in judgment or strategic lying. It doesn't necessarily indicate that people with higher belief are less likely to search for further information about their skill.
4. Women smile to indicate trust and warmth, while men smile to indicate confidence and lack of self-doubt.
5. People are tuned to be more honest, more conformist and more carefree when happy or in sunshine and light, and more dishonest, more creative and more analytical when depressed or in bad weather or in the dark.
6. People who are selfless and support others with time, money or emotions end up supporting themselves more than if they received equivalent support from someone else - backed up by mortality studies and satisfaction studies.
7. People unconsciously mimic those they like or are interested in, and mimicking somebody makes you like them more and be more attuned to them and others, and them feel the same way back towards you.
8. People think left and down correlates with smallness and up and right with bigness. Eye movements betray what number you're thinking of.
9. People are programmed to be happier when they think fast, because it's an indication of lack of hardship, even when there's no difference in hardship.
10. People grossly underestimate the trustworthiness of other people because when people trust others and are wrong, they get feedback, but when people don't trust others and are wrong, they usually don't.
11. The best predictors for somebody cheating on their partner are previous number of partners, level of impulsivity, happiness with current relationship and viability of other options. Predilection, impulsivity, context and alternatives matter in all relationships - business and personal.

1. Counterpoint - something different happening in the background of the major direction
2. Harmony vs Melody - something big with lots of smaller supporting efforts

1. You can say twice as much using half of the words/content. Being clear and concise helps your point.
2. Similarly, complex writing makes you look stupid. Keeping it simple makes you look smart.
3. Difficult names are associated with danger. The simpler the name, the safer it sounds.
4. The more fluent your speech, the more knowledgeable and intelligent you're considered, but when a speaker hesitates before a word, the word that comes out of his/her mouth is better remembered.
5. Don't stop the storyline for flash - let the flash advance your storyline. For example, car chases shouldn't interrupt an action movie, they should further it. Similarly, flashbacks and backstory shouldn't be a detour, they should carry the story forward.
6. Never write something where all there is is the flashiness, even if it is advancing the story in the end - there should be something going on at the same time. For example, a sex scene is boring, but following a sex scene while wondering when the woman's husband is going to discover them because you know he just arrived home is suspenseful. And it then it allows the "flash" sex scene to advance the story (see 5).

1. Communication and synchronization keeps you alive.
2. Mobility matters to your survival in the case of the unexpected.
3. Concentration in one area does more damage than spreading yourself out.
4. Attack the flank.
5. Sometimes, it's not the quality of the attack, it's the length and strength of the supply chain (think Rommel vs Patten in North Africa).

Political Theory, Law and Government
1. Man of System - people who want to control systems like chess pieces, not treating them like independent objects. Adam Smith has a wonderful quote about this.
2. Decentralization vs centralization
3. Bureaucracy, inefficiencies in responsibility designation and accountability
4. Property rights and legal protections
5. Equality of opportunity, and striving for justice
6. How does one relate equality of outcome with equality of opportunity, given unperceived differences in ability and effort and their distributions?

1 comment:

  1. It's a combination of Incentives (under Econ) and Tax-avoidance (under Finance), but the Laffer curve is something I've been thinking about lately in the context of all the tax bill debate