Important Information, Important Caveat

"between 210,000 and 440,000 patients each year who go to the hospital for care suffer some type of preventable harm that contributes to their death"

http://www.propublica.org/article/how-many-die-from-medical-mistakes-in-us-hospitals

Harm from medical care is a very big problem. Many hospitals are amazingly incompetent, failing to institute basic care guidelines that would save a lot of lives. If you have any choice about what hospital to go to, try to find information about the medical accident rate there. This will be difficult; hospitals always try to hide any information that could be used to judge their quality.

In general, medical care is both more dangerous and less effective than most people think. This new study is evidence that going to a hospital is more dangerous than we thought it was.

However, it does not mean that they 'kill' over 200,000 people. It does not even mean that all of those people would be alive today if they had done everything perfectly. It means that those people died, and somewhere in their medical records was a mistake that could have been bad. The number of people actually killed by these mistakes is likely to be a small percentage of the 210,000 to 440,000 guess. Even then, many of the mistakes were ones of omission, where the doctors missed something they should have caught. In such cases, staying away from medical care would clearly not have saved the life of the victim.

When making medical decisions for yourself or your loved ones, you need to compare two probabilities: the probability that getting medical care will harm you, and the probability that not getting any care will harm you. Many of the people who died from preventable medical errors may have died sooner if they never got any care.

But still, if you do not really need advanced medical care, it is usually wise to stay away from hospitals.

Whack Rant

I just started reading 'A Whack on the Side of the Head'. I had read that it was very good, so I borrowed it from a friend when I had the chance recently. But when I got to page 19, I became so disgusted with the author that I yelled "You Idiot!" at him and threw the book down in disgust.

What follows will be a rant. If it is like my previous rants, it will be heartfelt, entertaining, persuasive, compelling, and quite possibly wrong. You have been warned. But the rant was in my head, growing and taking form from the moment I threw the book down, and I have to let it out.

Here is the text that set me off, somewhat edited:

"To be more creative, all we need to do is "look at the same thing as everyone else", and then "think something different". Humans have been using their imaginations this way since the beginning.

...

The first person to look at bacterial mold and think "antibiotics" did this."

This is wrong. This is not how penicillin was discovered. One of the greatest public health miracles of the century did not happen because someone had a creative idea. What actually happened was that someone was running an experiment, noticed something strange about a plate of bacteria, and then started investigating further. Doing random things and observing the results, not premeditated creative thinking, was the real source of antibiotics.

His claim about antibiotics is wrong so thoroughly that it immediately and permanently discredited his entire thesis. He obviously does not know, or care to know, about the actual, real-life operation of a very important human invention. He has simply made up a narrative to fit his thesis. It is a breathtaking example of intellectual dishonesty. This is an unforgivable sin in someone trying to tell me how to change my thought process make good things happen.

I must assume that if he cares so little about truth and reason where I can catch him, he must be just as cavalier in the places I cannot catch him. I must assume that all of the rest of his examples and evidence are just as false, and therefore that any conclusions he has are false, wrong, and useless. It is clear that he has imagined a reality that fits his prejudice, rather than forming his ideas by observing reality.

Until I encountered this monstrous lie, I was prepared to accept the book's thesis. When I was kicked out of my acceptance, I immediately started thinking of other information that opposed the cult of 'creative thinking' Just like in the case of penicillin, progress does not come from thinking. It comes from tinkering, experimenting, observation, recording and sharing the results of random tests and observations.

The history of actual science, and our knowledge of how the human brain works, both tell us that thinking mainly exists to explain what has already happened. For most of human history, discoveries came first, and then the theories came later. People did things, got their hands dirty with the physical world, ran experiments and observed the results, tinkered and explored, and after all that was done, the creative thinkers spun theories to try to explain the results. Most of these theories were worthless, but some were right, and produced testable predictions, and then more experiments were used to figure out which theories were useful.

Thought is, to a first approximation, worthless. Humans have been thinking for thousands of years, producing vast volumes of philosophy and theology that are basically useless. Human life remained nasty, brutish, and short until we stumbled across the habit of running experiments and sharing the results.

If you want to accomplish something with your life, then the best advice is to stop thinking things and start doing new things. Try new activities, make new friends, and learn new skills. Humans spend way too much time just sitting around and thinking, and they accomplish almost nothing by doing so. Much 'thinking' is just a way to try to signal good qualities to people we want to impress.

Rich countries today are full of overeducated "creative thinkers" with no useful skills. If you talk about your creative thinking in a job interview, that is a strong signal that you have accomplished nothing in your life. People want results, and they want people who can deliver results, and results come from doing lots of different things and developing skills and good habits. Experience is far more important than thinking, and you will not get experience in anything useful by thinking about things.

I freely admit that I do not follow this advice. Overthinking things is a vice of mine. I even got lucky enough to land one of the few jobs where creative thinking is actually a useful skill. But I did not get the job because of my creativity. I got the job because I had produced a good job market paper; I had demonstrated that I had already accomplished something that was very similar to what they wanted me to do in the job.

The best businesses today do not rely on creativity. They rely on data and tests. They experiment with lots of things, observe the results, tweak things, and experiment again. Creativity is one small and almost unnecessary part of this process; it can tell what it might be interesting to test.

Creativity is mainly an asset in the bizarre, artificial, and manipulative world of the entertainment industry. Human do crave creativity in their entertainment. I argue that they crave it for exactly the same reason that they crave airbrushed images of sexually attractive people. It is something that we have evolved to see as a good signal. Creative people make good political allies because they can find persuasive ways of arguing that our side is right and their side is wrong, so the tribe should banish them and reward us.

Okay, the rant is winding down. I see no reason to try to keep it going or tidy it up. If this were a LessWrong group meeting, it would have started an interesting and creative discussion and I could start enjoying listening to and sharing the creative thoughts of potentially useful allies...

Ideal Hedge Fund

A friend asked me an interesting question today:

"If you could design an ideal, i.e. optimally efficient and proficient hedgefund, how would you go about it?  What features would make it unique?"

As I was answering, I realized that it would make a good blog post. So:

Short answer to question 1: The optimal hedge fund is a venture capital firm that either operates in an industry served by no other venture capital firms, or uses validated testing methods to judge the people they invest in.

Short answer to question 2: Actually pay your traders for alpha, by correcting their returns for both average market returns and the beta of whatever they are doing. None of this '2 and 20' nonsense that just rewards people for sitting on a lot of risky assets in a bull market.

Long answer: Asking an economist to design an ideal hedge fund is like asking an engineer to design an ideal motor. The structure, size, and organization depend in the intended use and a lot of other design constraints. If you are actually looking for a way to hedge risk, then it matters what risk you are trying to hedge against.

If by 'hedge fund' you mean a way to make excess returns by taking advantage of market inefficiencies, which I assume, then the important thing to remember is that the only way to beat the market is to be smarter than the market. That means being not only smarter than every individual trading in the market, but smarter than their entire combined intelligence as coordinated by prices and institutional knowledge.

It is basically impossible to be smarter than a large, liquid, well-analyzed market (unless you have access to Strong AI and nobody else does). You have to look for markets that are underanalyzed and illiquid, where highly specialized knowledge is needed to operate or that trained finance people have neglected so far. For example, the Harvard endowment was known for investing in things like timber land.

Of course, finding such markets is also a hard problem, because you have to be smarter or have better connections than everyone else attempting the task. However, you may have success by hiring subject matter experts with esoteric knowledge, and then spending the time and money to teach them about finance and markets. These experts should probably be experienced industry insiders rather than academic experts, although you will need a few people with technical and quantitative skills for the serious number-crunching.

Then you have to give them the right incentives. You have to make sure that you are not rewarding churn, but also not rewarding just sitting on assets. Actually pay your traders for alpha, by correcting their returns for both average market returns and the beta of whatever they are doing. This would be the unique internal organization of my ideally managed hedge fund.

I'd guess that the biggest market inefficiency in our economy today is that small companies are starved for investment because banks are trying to recover their balance sheets. They are being crazy strict with their small business loans, but in their defense, knowing which small businesses are worth investing in is a hard problem. You need good industry experience, good finance knowledge, and probably most important, good skills at reading people and judging their competence and character.

If you can get away with it, maybe you could use tests of various kinds, from personality tests to practical tests related to important business skills, to judge the loan applicants. That would be unique, and almost certainly more effective than standard interviewing techniques. However, be warned that the character traits that make good employees, like IQ and conscientiousness, are not the most important for entrepreneurial success. That seems to be based on perseverance, networking skills, and other things that valid testing instruments do not exist for.

If you invested in researching and validating such tests, then perhaps it could be a source of competitive advantage. I understand tech forms like Google do similar things for choosing employees. But the fact that nobody uses applicant aptitude tests to make venture capital investment decisions may suggest that there are legal problems. If your tests were discriminatory, you might get in trouble. That would require legal analysis.

Constructive Criticism

My boss just told me, "For a very intelligent person, you are really good at taking constructive criticism."

It seems obvious to me that I am very intelligent because I am good at learning from constructive criticism. Nobody is born knowing anything useful. Almost by definition, being an intelligent person means that, at some point in the past, you were good at taking constructive criticism from somebody. 'Learning from constructive criticism' is basically the definition of 'education'.

I walked into this job with the attitude that my bosses and the other experienced people were my teachers, and I was a student. I treat them like I treated my professors in grad school, respecting and deferring to their superior skills and experience while being unafraid to add my own knowledge, ask questions, and make suggestions.

I consider it right and proper that my draft documents will come back to me with dozens of comments and edits. I am the apprentice, and they are the masters. My work will inevitably have problems until I learn the approach, techniques, and presentation that the organization requires.

Even when the people editing and commenting on my documents are not the masters of my profession, they will always know something that I do not, or have a perspective that I need to consider. I can only learn from them, or learn how to communicate with them, if I respect and react well to their criticism.

But it is not just my intelligence that is boosted by my attitude to criticism. It actually helps my creativity as well. In the process of creation, I do not simply copy what has been done before. I do what I know is right, not what I think they want. I do not self-censor. This freedom and creativity comes because I know that my work will be criticized and edited as necessary, and I do not fear this. I am not afraid of being told to change. It is far better to see someone else squash my creativity into a box than to unconsciously internalize a stifling set of boundaries and limitations.

The key to this process is to sever any connection between my ego and the first draft of my work. Being told that I need to make big changes to my work is no threat to me, and I do not react with any defensiveness or hostility. I know that at the end of the process, my work will be very good. It will contain much of my originality and creativity and skill, while being molded in ways that make its readers better understand and accept it.

Mental Strangeness

Today I was confronted with something that made me realize how different I am from other people. It was a list of security questions for password resets. I assume that they were chosen because most people would easily be able to give a consistent answer to them. But most of the questions involved concepts, experiences, or mental processes that are alien or unimportant to me. I will run down the list:

"What is the first name of your best friend in high school?"

I do not have 'best friends'. This concept is alien to me. I have friends, and I do not rank them. There are three or four people with equal claim to 'best friend' in high school and my mind refuses to decide among them.

"Where would you most like to live?"

I have never thought about this and my answer to the question would likely change quite often. It seems a waste of time to dream and fantasize like this.

"What is the first name of your first boyfriend or girlfriend?"

The definition of the word 'girlfriend' is a very vague and fuzzy one, especially in the modern world. I am not sure if the first woman I kissed qualifies as a girlfriend or not.

"What city would you most like to visit?"

Again, I have never thought of this. There are lots of interesting places in the world that I might go to, but ranking them in terms of desire is never something I have done.

"What is the name of the street your childhood friend lived on?"

How am I supposed to decide which friend this refers to? I had lots of childhood friends. And I do not remember the street that any of them lived on, because my parents were driving.

"What is your favorite food?"

Again they ask me to rank things. I like lots of things, and my preferences change depending on factors like the season, what I have been eating recently, what I feel like, what the occasion is, and who I am with.

"What would you like your nickname to be?"

I hate nicknames. I want people to call me by my real name.

"Who is your favorite cartoon character?"

The media that I choose to consume at any time depends greatly on my mood and the situation. How can it make any sense to make arbitrary decisions like this? The word 'favorite' is inherently problematic for me. The very concept of picking a favorite seems both ill-defined and senseless in a wide variety of contexts. What is the point? Cartoons do not compete like sports teams. Am I declaring that I would always or usually choose to watch a cartoon with that character over any other? This is dumb. Am I declaring my affiliation to a certain social group of people that identify with the character? This also seems like a silly way of dividing and separating people.

"If you had the time and money to pursue any hobby, what would it be?"

Not only is this another ranking or favorite question, it presupposes that I desire to do something with my life that I am not currently doing. This seems like it would be an unpleasant state of existence.

"In what city did you meet your spouse/significant other?"

I do not currently have such a person.

"What was the model of the first car you drove?"

I have no idea which car my parents taught me to drive in. It was long ago and they no longer own it.

"What was the second state or country you lived in?"

Finally, a legitimate question about a well-defined fact of my life.

"What is the first name of the cousin who is closest to your age?"

I have lots of cousins close to my age, and I am not sure which one is the closest. I do not keep track of the ages of people with such precision.

"What was the first concert you attended?"

I have never been to a concert.

"What is your father's middle name?"

The second proper factual question.

"What is your maternal grandmother's first name?"

The third good question.

"What was/is your grandfather's occupation?"

I have two grandfathers, and each of them had several occupations.

Out of all these questions, I can only give a reliable and consistent answer to three of them. Most of the questions presupppose that I have generated arbitrary and irrelevant, but fixed and stable, ordered rankings about the things in my life. Many of the others assume that my life and that of my family followed some narrowly defined script.

One of the facts that people often do not understand about me is that I do not instinctively judge most of the things around me. Life happens. I experience things. I decide and act based on the situation and the constraints I face at the time. It seems a waste of mental energy to assign the things around me to arbitrary categories, or to compare and rank them. Judgments should be reserved for important things, and not the petty details of my personal existence.

Message to Yourself

If you could call yourself five years ago and had 30 seconds, what would you say?

Think about it. Consider several options. Think about your priorities. When you have made a decision, read on.

Most people seem to answer this question with advice for their personal life. This was one of my first instincts as well. This says a lot about the way that human instincts work. Our minds seem programmed to focus on petty personal things.

A slightly higher order of thought is to give yourself advice on buying stocks. This is much more useful than personal advice if your goal is to maximize your personal quality of life, but it is still a rather petty and selfish thing to do.

After a bit of thought, I started to consider what kind of tragedies I could have prevented if I had accurate information from the future. A lot of bad things have happened in the last five years. Crafting the ideal message then becomes a problem of identifying bad things that I could prevent if I had foreknowledge of them, guessing the probabality of preventing them, and multiplying to find the highest expected benefit for the world.

Probably the worst disaster of the past five years is the Syrian civil war. But I cannot think of any way that I could have prevented that catastrophe. Even if I was the actual president of the US, it is not clear that I could have stopped it. It is a messy diplomatic and geopolitical situation, and an unverifiable 'message form the future' will add nothing to efforts to prevent the disaster, even if that information turns out to be accurate.

The same is true for any number of natural disasters, famines, and diseases. I do not think that a random Economics PhD student could have any real chance of preventing such things, even with 30 seconds of accurate information. Presumably I would trust the message, but attempting to tell anyone else about the source of my knowledge would mark me as a madman.

However, there are much smaller tragedies that I might have had a good chance of preventing with a phone call to the right people at the right time. Despite the fact that the FBI gets millions of signals a day, and has an incredibly difficult job of choosing which ones to act on, a very clear warning about a specific person on a specific date would probably get a serious response. If I called in about a week before the Boston marathon bombings and gave the name of the perpetrator and said I knew he was planning an attack, they would probably start investigating.

Outbreaks of foodborne illness kill far more Americans each year than criminals and terrorists. If I was in a position of authority, then knowledge of things like the 2011 listeriosis outbreak from contaminated cantaloupe, or the 2012 meningitis outbreak from compounding pharmacy drugs, would allow me to save far more lives than knowledge of criminal acts that kill a half or a third as many people. But I do not think that a message from a random civilian would be able to start an investigation that would prevent these things.

So, my message to myself would be a brief description of the worst domestic crimes and attacks in the past five years, with dates and perpetrators. I have not researched this very thoroughly, but my first guess would be the Fort Hood shootings, the Sandy Hook shootings, and the Boston marathon bombings. Those are bad events that I could have a high probability of stopping if I had accurate information in advance of the attacks, and that information is probably the limit of what I could fit into 30 seconds.

4 Player Chess

Last night I won a game of four player chess by accident. I know that any of the other players could have crushed me in a two-player chess game, but the four-player version is a different game in interesting ways. For example, it is rarely a good idea to trade pieces, even if you can take a more valuable piece, because that means that your two other opponents are both in a better position.

The game is fascinating. Although the pieces move the same way, the strategic dynamics are very different. It is not possible to play 4-player chess the same way you play 2-player chess. If you try to think through future moves, you will be exhausted from the mental effort and your work will be for naught, because the board can change dramatically based on the actions of your opponents moving against each other. One player was paralyzed with indecision, not wanting to move anything, before giving up and turning his side over to someone else.

You have to play 4-player chess like you play Go. Instead of trying to be a chess computer, go with intuitions of life, breathing, influence, and position. From the start I did not try to think through things; I made bold moves to shake up the game and cause interesting things to happen. Very early on in the game I 'opened Pandora's Box' by moving a piece where it could have been costlessly captured by another player, pointing out that he had no incentive to capture the piece because it was threatening his opponents more than him.

Later, there were many times in the game where I had no obvious move, so I just shifted my pieces so that they had more options, threatened more squares, or were defended by more pieces. Someone else was asking me a question about fitting Vibram Fivefingers, so I was not paying much attention, and for a couple turns I was randomly putting pressure on the leading player by putting him in check. Then he congratulated me for a checkmate and game win. When I asked how that happened, he told me that I could move my queen right next to his king, where it was covered by my knight, and I had a bishop blocking the escape routes.

When you take someone's king, you control all the pieces left on that side. The player I could checkmate had already done this to another player. It had been a very swift reversal; it had appeared that he was aiming for one player but then he saw an opportunity to checkmate another player instead.

I suspect that the game could degenerate very quickly if you played to win rather than have fun. The optimal strategy is probably to 'turtle up' and build a fortress. Once the game is shaken up and moving quickly, it is fun, but you need a kind of cooperative 'play for the lulz' attitude to make that happen. It may not be a coincidence that I had been playing to have fun and not win, and ended up both winning and making the game fun.