Jurgis

Good article. I'm too lazy to create a WP account to ask the author the following question, but I'll throw it out here: So in winner-takes-all economy competitors cannot destroy the margins of AAPL/FB/MSFT/GOOGL. But that does not prevent competitors from trying. And when AAPL/FB/MSFT/GOOGL invest into trying to capture the market of FB/MSFT/GOOGL/AAPL, doesn't that drop their margins, since they can't win and they are just throwing money into the wind? I guess the answer could be that they spend only a small percentage of money trying to break into another winner's market. But is that really true? Didn't Google spend a lot on G+? Another answer could be that their winner margins pre-attacks-on-other-winners are so high that even the money wasted does not lower margins to past averages.

Hmm, rukawa, I am confused now. 8) Your last post is very pro-university, your post earlier is very anti-university system. Can you clarify? 8) Let me answer some of your points from previous post: The geniuses from the past who you like so much also had to find funding. OK, they were not exactly today's grants, but they had to find either government/state/government-institution or private rich people funding (or be rich themselves). How is that different? In some cases such funding was easier to obtain, in some cases harder. In some cases the constraints on the funding were less, in some cases way more onerous. You exaggerate there. So you don't believe in multiplier effect from graduate students doing research? You don't believe in graduate students doing worthwhile research? I do. I'll go with Google example again. Also, you seem to forget that graduate studies is studies. People who start at year one of graduate program don't know how to make good research yet (at least majority don't). At the end of graduate studies they do (majority do). So what you call "army" is really a way to teach young people to become researchers. And these young researchers will appear from where? From getting B.Sc.? You really believe that you can fund people to do research without them going to graduate school first? I disagree. Yes, some research results might be better if professors employed professionals instead of grad students. If you think that professors love having graduate students instead of professionals to do the work, then you are quite mistaken. It's headache for them too from that point of view. However, we are going back to the fact that you have to have institutions that train people on how to do research. And the current system is just a solution for that. Maybe you disagree with the whole concept of such institution. I believe it is needed and I believe it mostly works. Yeah, grad students are not necessarily well trained when they start working. Yeah, they don't necessarily know what they are doing. Yeah, when they become proficient, they graduate and leave. Yeah, these are the facts of life. Good professors and groups deal with that well. Bad ones don't. Some professors do hire professionals for some tasks. It is though expensive and thus limited to best professors and best institutions (MIT, Stanford, CMU). I think you contradict yourself when you talk about professors doing their work themselves and then talking about them hiring professionals to do things. Professors doing the work themselves would just mean less progress and less results. You may think that's good. I don't agree. Professors hiring professionals would make things way more expensive for whoever is paying for the research. If money is available, then I'm all for it though. BTW, what you argue for somewhat exists too, though in limited quantity. There are government research labs where researchers work themselves with some professionals on various research projects without student "army". There are private and industrial research labs. You may not like these for other reasons, but they also mostly follow the model that you like: "researchers working themselves with some professionals on various research projects". So we do have alternatives. Finally, you have a lot of emotions against the system. I guess that you had bad experience with it. I'm sorry for that. I've had great experience with it and although I have not worked at the university after doing my degree, I still interact with people from there and have a lot of respect for them. It is not ideal system. It has issues. But I disagree that it's "utterly without logic and reason. Its inefficient, corrupt and stupidly run." Good luck

Yeah, sure. Google for example. Oh wait.

I agree with Liberty. Sample bias and recency bias (you know bad apple examples from now, but you don't know the bad apple examples from Newton's time... oh wait didn't he and his students fight with Leibniz like heck? https://en.wikipedia.org/wiki/Leibniz%E2%80%93Newton_calculus_controversy ) There are tons of great researchers nowadays who do it for love of science. There are great researchers who achieve spectacular results possibly comparable to greats in the past. There are great professors that teach and guide students as well if not better than the past ones did. However, yeah, there is increasing specialization - the "easy" knowledge is all discovered and the new achievements require more and more esoteric specialization and understanding of very limited domains to great depth. It is no longer possible to be polymath to great extent. Some people still are somewhat, but not a lot. Sometimes there's also a "new" field in which you can discover somewhat "easy" knowledge like computer science in 1940-1970s or so. But that's somewhat rare.

What KJP said. Edit: and even if you do something like 60/40 or 70/30, you still can argue that 60-or-70 should be in-index or beating-index.

Since some of the posts went towards medicine, I'll post the following article that I found rather interesting: https://www.theatlantic.com/health/archive/2017/02/when-evidence-says-no-but-doctors-say-yes/517368/ IMO it's a well written article and possibly useful to some people who commented above. I appreciate comments of tripleoptician and lschmidt from the field. This article is only partially related to the original topic of this thread.

False. Even highly survivor-biased polls on CoBF, which is stacked by the best-of-the-best, does not show that it isn't hard to beat the index.

"Expert" is not just appeal to authority. Unfortunately, like nafregnum said, science is now so complex that somebody outside the field cannot understand all the data, all the studies, the variables, the possible interactions of variables, the statistics used, and the significance of such. What I will give you is that even within the field not everyone understands this and/or can summarize the data, evidence, studies, point out the questionable spots, etc. But for people outside the field it is simply impossible (unless you are a genius and/or spend couple of years studying the field) to do that. So, yeah, you can either rely on the accepted majority opinion based on long term studies and their results. Or you can decide that you will take the opinion of couple minority studies and possible effects of data mismanagement (which don't necessarily even support minority hypothesis). Or you can have a weighted opinion. IMO someone like Liberty has a weighted opinion. Someone like your original article doesn't. If you still believe this is "appeal to authority", so be it. If you still believe that you can have a better weighting about the pro/con/tampered/non-tampered studies without being inside the field and understanding it thoroughly than relying way more on the established long term expert opinion, then I disagree with you. Good luck

I won't. First of all neither you nor I are experts in medical studies. So even if I posted 99 studies, there won't be any productive discussion afterwards. Second, your phrase drastic conclusions immediately shows your existing prejudice against the scientific studies. Your attitude is to question an established long term results based on a couple samples that are questionable. While this might be productive if you were an expert in the field, for non-expert it shows a prejudice against the experts in the field.

Before I posted it, I checked some of the other articles from the homepage. I had the same response that you did. But I thought to myself: If the article itself is thoughtful, has legitimate points, and legitimate sources (I checked most of the sources myself), then it should be posted. The section on vaccines is what first made me question the article's authenticity: Sounds a little funky, right? Well let's check the source: http://morganverkamp.com/statement-of-william-w-thompson-ph-d-regarding-the-2004-article-examining-the-possibility-of-a-relationship-between-mmr-vaccine-and-autism/ So the source appears authentic, if you ask me. The source is likely authentic. The Wakingscience website though amps up the source into single sided "scientists cooked results, vaccines cause autism, etc." diatribe. Even if all sources of the article are authentic, reliable and not exaggerated (and some of them clearly are), there is still a huge anti-science slant by not reporting anything from the other side. Let's put it that way: if there's 98% of studies that did not cook data and reported X, 1% that cooked data and reported X, 1% that reported "not X", what would be your conclusion about X if you read an article that only mentioned the 1% that cooked data and the 1% that reported "not X", but said nothing about 98% of other studies? Would your conclusion be valid, interesting and justified?

I think your last conclusion is closest to the truth. I highlighted the key sentence above. I am waiting for Elon Musk'y integration of humans and cloud-AI, where I could run a comprehensive analysis of any claim, e.g. "aspirin reduces heart attacks while having no negative side effects", with all publicly-available data within couple minutes and have a reasoned statistical analysis (which I can still agree or disagree with). Then we might actually have a real current-fact-and-theory-level understanding and explanation of the world and any claims we hear. Yeah, this will still have drawbacks (e.g. you might not have facts, the past studies might be crappy, etc.), but it will move towards data-based claims/understanding. Some of responses you get in your experiment might just be politeness though. I might say something positive to a person who claims that "salt purges your organism from mercury" while thinking that he's total moron. 8)

Yeah, I like his writing and most of his buys.

Anybody has opinions about his income fund? Looks like a lot of risky bonds... but is he good in investing into them? I guess I should look at past reports on why the fund went down in 2014-2015. Probably energy bonds?..

Like the one that states: Emphasis mine. ::)

I agree with most of what oddballstocks said. Yes, there are issues with scientific studies. There are issues with reproducibility. There are issues with publish-or-perish. There are issues with incentives. On the other hand, most of these issues are not as bad, as pervasive or as broken as anti-science critics imply. Even in the medical field, look at the cancer treating advances in last 30 years or so. Look at HIV treating advances. Look at Hep C treatment. Yeah, there are crappy studies and marginal or not effective cancer drugs approved (and then possibly withdrawn). But that overall does not negate the huge life extension provided by treatments in last 30 years or so. Even with all the influence of pharma companies, there are people looking at study results, doing meta studies and pushing better analysis and better future studies. The whole reproducibility issue makes people (at least some people) construct more reproducible studies. Criticism of placebo-level-only results makes people make better placebo-controlled studies. On the third hand, how many of the people who criticize the studies have tried to make a large, placebo-controlled, double-blind, long-term medical study. These are horrendously difficult. That does not excuse all the simplifications, but the simplifications are not just because researchers are lazy, ignorant, on-the-pharma-payroll, etc. (E.g. because of ethics, you can't have control group on placebo if there are drugs that somewhat-perhaps-work and control group may die because you did not give them these drugs... ) On the fourth hand, and I mentioned this in another thread: a lot of devil is in the domain-specific details. So reading a popular-oriented articles about issues in field X is like watching a movie: some things are real, some things are way oversimplified, some things are distorted, some things are flat wrong. The problem I have with OP's article is that it is completely one sided. LC claims it's "well-sourced". But really it presents one side only and gives zero opportunity for the science side to answer any of the issues raised. Edit: I posted this before reading Liberty's reply. Liberty's pointers to other articles on that site completely shoots whatever credibility it had before. Yeah, I was right: it's a site for anti-science kooks.

So much one sided anti-science information. Lots of local anti-science kooks will be deliriously happy.

Right. And because of what frommi posted, markets may be somewhat efficient, i.e. it's hard to outperform by picking any horse - good or bad - without extra insight. Obviously there are exceptions as some people on this board and elsewhere demonstrate.

This is complete nonsense. Excusing Trump's election by "PC drove people to the edge" is utter bull crap. At least if you said that white male entitlement drove people to vote for Trump, you might be a bit closer to truth. We should not excuse misogyny, racism, homophobia, nationalism because their restriction supposedly drove someone to the edge. This is what the alt right racists want us to do. We won't surrender. We will not go back to the time when it was OK to denigrate people based on their gender, race, skin color or nationality. And BTW this forum is becoming a swamp of alt right demagogues because of no moderation.

What you are saying is that alpha is zero sum game (in closed system). This is true. What do you mean by "closed system" by the way? "Closed system" means that we are only measuring effects within the system itself, there are no leakages outside, and/or they are not included in the measurement. E.g. if you allowed for investors to hold non-index securities and/or cash and measured outperformance including these, then the zero sum game does not apply. Or more precisely it only applies once you make the index that includes these formerly-non-index-securities-cash/etc.

What you are saying is that alpha is zero sum game (in closed system). This is true. However, this in no way relates to the claim that "on a dollar weighted basis half of investors have to outperform an index." Not sure if you wanted to relate to that claim or not. 8) Maybe the claim "on a dollar weighted basis half of investors have to outperform an index." was supposed to express that alpha is zero sum game. Maybe it was supposed to say: "in dollars outperformers outperform by the same sum as the underperformers underperform". But that's not what it said. 8) Here's the quote from the Sharpe paper that proved it: "Over any specified time period, the market return will be a weighted average of the returns on the securities within the market, using beginning market values as weights. Each passive manager will obtain precisely the market return, before costs. From this, it follows (as the night from the day) that the return on the average actively managed dollar must equal the market return. Why? Because the market return must equal a weighted average of the returns on the passive and active segments of the market. If the first two returns are the same, the third must be also." Tbh I haven't completely wrapped my head around it, but i'm not sure i'd go around disagreeing with Sharpe and Mauboussin, who are both miles smarter than I am. I think my single transaction example works pretty well for me. If that's replicated for each transaction then on a dollar weighted basis (as the night from the day) half would outperform and half would underperform before fees I don't think we are disagreeing at all.

I'm not associated with this. Just came through my infoporn feed and I thought it was fun interesting opportunity: Columbia University Post-Doctoral Research Scientist, Machine Learning in Wealth Management http://jobs.acm.org/jobseeker/job/32715897/?utm_source=JobFlash&utm_medium=Email&utm_content=JobPosting&utm_campaign=JobFlash-3%2B01%2C%2B2017 "There is also an opportunity to be involved in a start-up company arising from the research of the Center." 8)

Interesting that this topic appears on the day Youtube TV is announced ( https://www.cnet.com/products/youtube-tv/preview/ ) and first four posts does not even reference it. 8) My take - as a coattaily Liberty empire investor - is that there's way too much content being produced. So although I hold DISCA and LGFB, I really should shit more to the cable cos (LBRDA, LBTYA, LILA). Actually that's my weighting already.

Yeah, it's an old joke. Heard it 20 years ago or so and likely it was an old joke even then. It was more culturally-OK then perhaps. Sometimes Buffett is just not very sensitive when he tells his jokes. ::)

What you are saying is that alpha is zero sum game (in closed system). This is true. However, this in no way relates to the claim that "on a dollar weighted basis half of investors have to outperform an index." Not sure if you wanted to relate to that claim or not. 8) Maybe the claim "on a dollar weighted basis half of investors have to outperform an index." was supposed to express that alpha is zero sum game. Maybe it was supposed to say: "in dollars outperformers outperform by the same sum as the underperformers underperform". But that's not what it said. 8)

Mhm. :) Although topofeaturellc is trying to make this more complicated by talking about investors holding the shares. Still what you said is true and is perhaps better example/explanation than mine was.