Basic error is that you think you can predict fully random process. You develop a model that seems to work excellent until it doesn't.
It doesn't matter whether it's sophisticated algorithm running on big iron or hunch running on wetware. You will fail because nobody knows the future. Trying to reliably squeeze money out of the market by better predicting its behavior is like trying to build perpetum mobile by trying more and more convoluted designs.
There is money that can be reliably extracted from investing though. If you can trade faster than others, you can drain some money from them. That's why algorithmic trading yields returns. Your strength comes not from predicting future better, but from acting and reacting faster.
Certainly you need to protect against over fitting your models, and you need to consider tail events.
But to say that the entire market is fully random is to claim that the efficient market hypothesis holds, which it clearly doesn't - not even in weak form. Anyone who wants to claim that the EMH is true needs to explain the equity premium puzzle, the success of value investing and the persistent outperformance of momentum investing - among many other price anomalies that I could name.
Overfitting your models? If you fail to predict an earthquake or a tornado do you blame it on overfitting your model? I don't think so. You just humbly say that you don't have good enough model, enough computing power, and what's most important not enough data.
Value investing is just buying things that you think are under-priced. It's successful if you are right and you are faster then other people to notice that, or you are actually wrong but lots of people also make same mistake but later than you. Or if you are Warren Buffet then other people think you must be right and buy the same thing making you even more rich.
As for momentum investing it's easy to just read wikipedia entry on that. It's not even established that this effect exists and even if so it's offset by additional risk undertaken by adhering to this rigid strategy.
I'm not sure how Equity premium puzzle even relates to efficient market hypothesis.
People clearly don't believe that efficient market hypothesis holds but same can be said about any recurring lottery. People keep inventing systems for picking lottery numbers, and claim that some of them work for some of them for some time. Which they do, until they don't.
When you are investing you don't play against the environment. In fact environment is favorable to you. You just compete with other players to rip off other players. It's so full of chaotic feedbacks that you can't predict anything. I mean, you can, until you can't. It's just like gambling. You can always win at fair gambling, just bet more if you loose. You'll be winning until you loose so many times in a row that you loose everything.
> But to say that the entire market is fully random is to claim that the efficient market hypothesis holds, which it clearly doesn't - not even in weak form.
Of course it does. All evidence supports the Efficient Market Hypothesis (EMH), and note my use of the word "evidence". A contradiction would have to be a series of funds that applied a given, predefined (internally) strategy and (very important) didn't reveal its picks to the public (to avoid the announcement effect). Such an experiment hasn't been done, and those cases in which a specific strategy was applied have been universal failures.
Let's say for the sake of argument that there is a surefire winning strategy as is so often claimed in worthless investment books. If it were to be recognized and then applied, it would become the new market norm, everyone would apply it, and its differential value would evaporate overnight. That hypothetical experiment merely supports the EMH.
> Anyone who wants to claim that the EMH is true needs to explain the equity premium puzzle, the success of value investing and the persistent outperformance of momentum investing - among many other price anomalies that I could name.
Those result from a combination of the announcement effect and insider trading. Berkshire-Hathaway is successful because people expect it to be successful, and who invest in the same stocks when they're announced. It's a self-fulfilling prophecy.
The only fair, scientific test would be a strategy that is applied in secret, not in public like Berkshire-Hathaway, and in a way that doesn't tweak the market itself by way of large purchases that create their own following of fanboys. In other words, a pure computer model that is defined in advance, not tweaked enroute, that doesn't actually invest but only models hypothetical investments, and that shows its value over other systems in perpetuity. There is no such thing, and the EMH is true.
Remember the Null Hypothesis, the basis of modern scientific thinking. The Null Hypothesis assumes something without evidence is false. All evidence supports the EMH.
It's tough to know where to begin here. To refute the EMH you don't need a 'surefire winning strategy'. You only need a strategy that is profitable in expectation. And it doesn't need to be laid down in advance - all sources of expected return have a half life, and strategies need to be retired when they stop working. That necessitates continuous research.
Also, you seem to defeat your own argument. The strong form EMH says that excess returns are not possible, even with insider trading - which you admit can lead to outperformance. The weak form EMH would prohibit excess returns from the announcement effect, but you admit that leads to return predictability as well. Which is it - does the EMH hold, or is the announcement effect real?
Your arguments for why momentum, value and equity investing work also don't stand up. There's nothing to be "announced" in those strategies - you could apply them in a vacuum where you didn't know anything except prices and fundamentals. Similarly no insider knowledge is required. So why do you say that they are due to the announcement effect and insider trading?
You also misunderstand the scientific method - the failure of an experiment to refute the null hypothesis does not lend credence to the null hypothesis! At most it tells you that the null is consistent with the observations. Moreover, you talk as if the Null Hypothesis is a particular statement, which it is not. There's a great XKCD comic I'd refer you to if I wasn't on my phone!
If you want such an experiment, I humbly invite you to consider the past and future returns of the fund I work for, which is 100% model-driven. Although obviously we trade based on our predictions - because at the end of the day we're in this to make money, not to write papers or win internet arguments.
> To refute the EMH you don't need a 'surefire winning strategy'. You only need a strategy that is profitable in expectation.
Yes, and, in spite of its tremendous theoretical importance, there is no such thing.
> And it doesn't need to be laid down in advance ...
Is this a scientific conversation? Yes? Then yes, it does. Any number of bogus studies can prove their claims if one can change the study's claims as the evidence comes in.
> The strong form EMH says that excess returns are not possible, even with insider trading ...
Ah, I see -- redefine the EMH, then prove the falsity of the redefinition. The EMH says no such thing.
> The weak form EMH would prohibit excess returns from the announcement effect ...
Also false. You aren't discussing the EMH, you're making it up as you go along.
> There's nothing to be "announced" in those strategies - you could apply them in a vacuum where you didn't know anything except prices and fundamentals. Similarly no insider knowledge is required. So why do you say that they are due to the announcement effect and insider trading?
That's easy to answer -- I never said any such thing, anywhere. You're arguing against a position no one has taken.
I did say that an legitimate test needs to be conducted in private, with no actual investments -- a computer model defined in advance, that cannot be changed as the study proceeds, and that by definition cannot influence the real market being studied. There are no such successful models. The EMH appears to be true.
> If you want such an experiment, I humbly invite you to consider the past and future returns of the fund I work for ...
Ah, so you have a dog in this fight. You're not an impartial participant, striving for objectivity. That means I'm wasting my time talking to you -- you're not a scientist, you're a salesman.
Whatever the merits of your arguments, your ad hominem at the end, as well as the general tone of your comment, undermine your position almost completely.
What ad hominem was that? I correctly identified my correspondent as a salesman, not a scientist. It's a simple statement of fact. He thinks one data point counts as evidence, which means he isn't a scientist, and he is objectively in the business of talking people into investing in his fund -- therefore he is a salesman.
Simple straight-up facts. And it really is pointless trying to discuss these issues with someone whose livelihood depends on the myth that the market can be tweaked in any way we please.
Perhaps you're like me a decade ago and you don't even realise when you're insulting someone (I was like that). If so, I urge you to realise that just because you don't mean to be insulting doesn't mean you aren't.
The last line of your comment was obviously insulting and your comment has been downvoted by others as a result. That's an evident observation. Take it on board instead of disputing it, and if you really did not mean any insult, then think about how you could have phrased that line (and perhaps the rest of your comment) in a way that would not come across as insulting.
Learning to not involuntarily insult people is pretty useful.
> Perhaps you're like me a decade ago and you don't even realise when you're insulting someone ...
You have something against salesmen? I described this person as a salesman, after he identified himself as a salesman. Which part of this aren't you getting? He has a vested interest in the position he was putting forth. That fact, the fact that his livelihood depends on the continued belief in the predictability and malleability of equities trading, is germane to the conversation. It prevents objective evaluation of evidence.
> ... in a way that would not come across as insulting.
Not every salesman is Willy Loman, and the insult you saw is solely in your mind.
He said: "I humbly invite you to consider the past and future returns of the fund I work for ..."
By referring to "future returns", he went beyond any rational position, and he directly contradicts the position taken by the SEC on this issue (brokers should never make claims about future performance).
He is a salesman. Talking to him about this issue is a waste of time. Those are the facts.
No, those are not the facts. That is your interpretation, and it is worded in your vocabulary. That vocabulary means something to you but may mean something else to someone else.
The wiser way to behave here is to realise that your interpretation may or may not be true, since you are working from a highly incomplete data set and therefore tone down the use of vocabulary that may have multiple meanings and instead focus on vocabulary that is less likely to be interpreted as an insult.
For example, you might have said:
> I notice from one of your statements that you seem to be working for a fund yourself. Do you think that this impacts your view of this problem?
Instead of:
> Ah, so you have a dog in this fight. You're not an impartial participant, striving for objectivity. That means I'm wasting my time talking to you -- you're not a scientist, you're a salesman.
All the bits in italics (yes, that's most of your comment) are potentially insulting to the person you're replying to. The first one is a mild character attack, the second one is a direct insult, and the third one is easily perceived as a direct character attack, especially in the context of the other two. Overall, this sentence is the comment equivalent of sneering at someone at a well-to-do party, and loudly dismissing them as a damn fool. If you're the host of the party or someone notable, you may get away with this sort of behaviour, but as a guest, you will be the one led to the door.
The restatement above makes the exact same point, casts the exact same aspersions on the views of the person you're replying to, but does so in a way that is not perceived as an insult.
> That is your interpretation, and it is worded in your vocabulary. That vocabulary means something to you but may mean something else to someone else.
This is a scientific discussion, not one about vocabulary or a post-modern view of reality.
> All the bits in italics (yes, that's most of your comment) are potentially insulting to the person you're replying to.
That's the way it is in science. Science is not diplomacy. And the WSJ Dartboard Contest went on insulting the same people in the same way for years, but without lifting the veil of ignorance in which these people live, as the above exchange proves.
> By referring to "future returns", he went beyond any rational position, and he directly contradicts the position taken by the SEC on this issue.
You asked for an example of a strategy which could reliably beat the market. I offered up the example of the fund I work for. Do you really think that this is me being a salesman?
Here are "the facts":
1. I am not, nor have I ever been, employed in sales. I work in research. I have zero client contact.
2. If I did want to raise investment for the fund, I wouldn't come to an obscure discussion thread on Hacker News to do it. Apart from anything else, most people can't afford the minimum investment.
3. It's an SEC requirement to include something like "past performance does not guarantee future results" on advertisements for mutual funds (NB this isn't an advertisement, and I work for a hedge fund, not a mutual fund). Strictly speaking that's true - nothing guarantees future results. But let's be Bayesian about it - would you rather put your money in the hands of a manager who's returned -10% for the last five years, or one who's returned +10%? [0]
Finally, if you think it's a waste of time talking to me, you presumably think it's a waste of time talking to anyone who works for a hedge fund (or any other kind of fund).
What you're saying is "The opinions of those who have the most experience on this issue, who spend 60+ hours a week thinking about this, and who put their own reputation and investments on the line daily -- those opinions are the ones least worth listening to."
I appreciate your point about conflict of interest. It's true that if everyone believed in the EMH, I wouldn't have the job I do now. But to summarily discard the point of view of everyone who works in finance seems a little bizarre. It's as if I said that it's not worth talking to software developers about programming, since if no one believed it was worth buying software, they wouldn't have a job.
[0]It's a trick question. If you believe in the EMH, your prior on the expected excess return on any strategy is 0% (or slightly negative to account for trading costs) and no amount of data will change that. You should be perfectly indifferent to these two money managers. But are you? What if they had 20 years of history? What if they had 100 years? At what point do you start to revise your opinion about the EMH?
> You asked for an example of a strategy which could reliably beat the market.
I never said any such thing. I asked for scientific evidence that the EMH is false. There is no scientific evidence that the EMH is false -- none whatsoever.
A single data point is not science. Your experience is not science. Your claim about future performance is certainly not science. TO have scientific value, it must be weighed against the entire population of stockbrokers.
> What you're saying is "The opinions of those who have the most experience on this issue, who spend 60+ hours a week thinking about this, and who put their own reputation and investments on the line daily -- those opinions are the ones least worth listening to."
That is exactly right -- every word. People who are closest to this issue, who have a vested interest in a given outlook, are the least reliable sources of evidence. And this is Science 101.
Do we ask murderers their opinions on sentencing guidelines? No, we don't -- their views aren't objective. By the same token, we don't ask stockbrokers their opinions on the theoretical underpinnings of the EMH, for the same reason -- their views aren't objective.
Your position is common among people who don't understand science -- "There is a theory that the equities market cannot be played, that the average return, reflected for example in the DJIA, cannot be improved on in the long term and over all investors and funds. But my experience disproves it -- I work for a fund that is at the right-hand side of the standard distribution, therefore the EMH is false." This is laughably absurd.
No one who understand statistics or science can miss the fundamental logical error you're making. I wonder -- do you by chance live in Lake Wobegon, where all the children are above average?
> At what point do you start to revise your opinion about the EMH?
First, it's not an opinion, it's based on scientific evidence. Second, it would be easy to disprove the EMH in principle, and may have tried but failed. Are you not aware of the Wall Street Journal Dartboard Contest? It is an excellent example of a pragmatic "put up or shut up" challenge to stockbrokers such as yourself, and it shows that there is no basis for the claim that managed accounts are worth their costs, or that the EMH is false.
> But to summarily discard the point of view of everyone who works in finance seems a little bizarre.
First, the views of individuals must be discarded, for reasons that should be obvious. Second, the experience of all stockbrokers as a population is the only rational basis for objective scientific evidence. Also, if the point is to find out which outcome is true in a study of human behavior, the least reliable source of evidence are the opinions of individuals who practice what's being studied. One would instead examine the actual outcomes, not the views of those outcomes expressed by the practitioners.
Let me spell this out for you. In science, your personal experience means precisely nothing. In science the average experience of all people in your position might lead to a useful result, depending on how careful the data was collected and processed.
Case in point. J. B. Rhine of Duke University believed he had uncovered evidence of ESP -- not because of a single data point, but because of many experiments over years. His data showed a persistent, small bias in favor of the precognition idea.
After Rhine passed on, others examined his records and discovered that he had been systematically dropping certain experimental outcomes, on the ground that they were so bad that they couldn't possibly be valid. In other words, Rhine didn't understand statistics, in which, given sufficient trial runs, an occasional very bad result is likely.
The excluded results were reinstated, and Rhine's ESP claim evaporated. True story. Now, you are saying that your personal experience disproves the EMH, or deserves to be counted as evidence against the EMH. But you are wrong -- this is not how science works. If all people in your position experienced the same thing, that would be a different story, but as is well-established, the average performance of stockbrokers is average, not exceptional (the dartboard contest proved that).
> It's as if I said that it's not worth talking to software developers about programming ...
But that's true! If the point is to discover how many lines of code contain bugs, the last people you would want to ask are those who wrote the code. You would instead work out an objective and preferably automated way to detect bugs, to avoid the trap of depending on the opinions of people in a given line of work. In study after study, 80% of drivers think they're above average, 94% of college professors, and so forth. It's mass delusion, and it's why human studies have to be performed so carefully.
Title: "94% of College Professors are above average"
Given the above well-established bias, the Wall Street Journal Dartboard Contest was meant to avoid asking stockbrokers what they thought, and gather evidence in a more objective way -- by comparing stockbroker performance against a randomly chosen pick of stocks. The dartboard's outcome was roughly equal to that of the brokers (within experimental error bounds), which means people who pay brokerage fees are being cheated.
Only in academic circles does EMH stay true. Time and time again, reality has shown that EMH is not applicable, or exists in a very weak form. One need only look to the LBO bubble and the dotcom bubble to see counter-examples of EMH.
In a conversation like this, rhetoric is cheap, and there is no evidence for your position. All evidence suggests that the EMH is true. Prove this wrong -- identify a surefire winning stock picking strategy that is immune to the announcement effect and insider trading, that reliably puts its practitioners ahead of market indices. There is no such thing.
> One need only look to the LBO bubble and the dotcom bubble to see counter-examples of EMH.
Those are not counterexamples, and they bear no relevance to the present topic. Anyone can misbehave in spectacular ways, but those misbehaviors don't falsify the EMH, any more than an airplane crash falsifies theories of aerodynamics.
First, from Wikipedia: "Empirical analyses have consistently found problems with the efficient-market hypothesis, the most consistent being that stocks with low price to earnings (and similarly, low price to cash-flow or book value) outperform other stocks." At the very least, accept that it's not such a clear case as you present it to be.
But then, what about HFT arbitrage, low P/E ratios, choosing underperforming companies to take advantage of regression to the mean, Piotroski's F-Score, Graham-style value investing, bait and switch dumping, swing trading, the very real problem of pricing goodwill and non-material resources...? If not undermining weak EMH, at least they cast doubts over the strong forms. Or they should. Do also read John Carter's Mastering the Trade for some examples on how market inefficiencies can be exploited reliably.
Supposedly (seen in Wikipedia, so I'm not that confident) market crashes are interpreted to be impossible under strong EMH (and yeah, the bubbles would definitely falsify them), but are allowed in weak EMH to be statistical events of a random process. Now, I've heard an interesting and convincing argument by Steve Keen[1] that the market is not Gaussian noise; it's a fractal process with dimension ~ 1.7 (Gaussian's dimension is 2) whose crashes obey a power law distribution, very different from what a random walk would generate. That falsifies the weak form.
> Prove this wrong -- identify a surefire winning stock picking strategy that is immune to the announcement effect and insider trading, that reliably puts its practitioners ahead of market indices.
Falsifying EMH does not require such a thing, no more than proving relativity requires observing time dilation at all velocities. This is an onerous, diabolical burden of proof.
> Anyone can misbehave in spectacular ways, but those misbehaviors don't falsify the EMH, any more than an airplane crash falsifies theories of aerodynamics.
If you had periodic epidemics of plane crashes that were more frequent than expected, sure, you should definitely suspect something's wrong.
[1] Yeah, some things he says are somewhat crackpotty, but he's definitely not to be completely dismissed beforehand. I refer to his Behavioural Finance Lectures, published in Youtube.
>> Prove this wrong -- identify a surefire winning stock picking strategy that is immune to the announcement effect and insider trading, that reliably puts its practitioners ahead of market indices.
> Falsifying EMH does not require such a thing ...
As a matter of fact, yes, it does. The only acceptable evidence against the EMH is a performance that the EMH cannot accommodate -- a successful strategy for reading and exploiting the market in a way that performs better than market indices. And not just sometimes, but all the time, reliably.
Science requires evidence. Evidence in this case means producing results above the norms set by the market averages, and not just sometimes, but all the time. Anything else is empty philosophizing.
Performances below the norms can be explained in any number of ways, but above the norms is a different story. And by "above the norms" I don't mean 1/2 the time or 1/2 the practitioners -- that's assured by chance factors.
All this talk about the EMH being flawed, not being the whole story, falls down before the fact that no one can invent a strategy to improve on long-term market averages, and that says it all. It is the ultimate evidence in favor of the EMH.
If no one can think of a way to read the market to his advantage, better than average performance, without engaging in insider trading or reliance on the announcement effect, then the default assumption is supported.
> This is an onerous, diabolical burden of proof.
Be that as it may, it is an essential standard of proof. Nothing else matters -- no amount of philosophizing can stand in for actual results.
The expert gets up and says, "The EMH is flawed, meaningless, unscientific, wrong, etc., etc.." Someone from the audience says "All right -- prove it. Produce a strategy that beats the market norms." The expert is naturally taken aback -- "Are you serious? What about the EMH? What about the fact that every single strategy that has ever been tried, has failed in the long term?"
This is like people in a bar talking about sailing. Any number of ideas about sailing are aired over intoxicating beverages, and it's all very interesting, but since no one goes out on the water, it's completely meaningless.
Proving that EMH is false requires revelation of a profitable method or at least a statistically significant edge, which then makes it vulnerable to getting exploited away. It's actually more useful if the public believes in EMH and in traders' best interest to keep their edges secret.
> Proving that EMH is false requires revelation of a profitable method or at least a statistically significant edge, which then makes it vulnerable to getting exploited away.
Yes, true, and such an outcome ironically supports the EMH.
> It's actually more useful if the public believes in EMH and in traders' best interest to keep their edges secret.
Actually, it is much, much more advantageous for traders to pretend that there is a surefire winning strategy that only they know about. The problem with that plan is the fact that there is no such strategy.
How do I know this? If there really was a surefire winning strategy, a trader who controlled it could (and would want to) do without any public involvement, and invest his own funds in private until he became absurdly rich.
That's the obvious way to play it -- why share the proceeds, and why risk exposing the method, by dealing with gullible, ungrateful clients? The answer is that those gullible, ungrateful clients are what makes the present system work, not some secret sauce that falsifies the EMH.
We are talking here at such high level abstractions and with stack of presuppositions that it's difficult for me to figure out just where to even begin explaining my point and i question if i even should.
If one doesn't know of a 'sure fire winning strategy' (== strong form inefficient market hypothesis ?) ( or just statistically useful may be good enough ) one should not participate in the market in most forms ( which market, participate how ). I've heard it compared to a poker game where if you don't know who the sucker is it's probably you.
The general public is misinformed about the markets and gets consistently fleeced yet keeps getting pulled in by dreams of easy money - akin to a lottery.
Markets have defined purposes (theoretical, practical, legal) and serve them accordingly. Enriching 'gullible, ungrateful clients' is not the primary purpose. EMH is a useful explanation to the clients why they got fleeced, and i don't want to disprove it.
It takes a lot of work to figure out how the lottery works and how to consistently game it. It's just a job.
> ... or just statistically useful may be good enough ...
If "statistically useful" means reliably above market averages year after year, then that would constitute a surefire winning strategy. There is no such thing -- if this were not so, we all would know about it by now.
A tiny advantage -- a legal one, not dependent on insider trading or any other questionable premise -- however small, could be exploited to enrich the player who used it.
But there is no such thing.
> The general public is misinformed about the markets and gets consistently fleeced yet keeps getting pulled in by dreams of easy money - akin to a lottery.
Yes, and the key to all these misleading strategies is ignorance of the EMH.
Evidence for the EMH lies in the fact that, in the long term, buy & hold investors do better than players -- in the long term, on average, if all results are tallied objectively.
> It takes a lot of work to figure out how the lottery works and how to consistently game it.
Again, there is no such thing. If there were, businesses would refuse to use equities to raise capital, knowing they would be fleeced by the players. The only reason equities are a reasonable way to capitalize business is because it's a fair approximation of an efficient market.
For the life of me, I can't figure out why people find this so hard to understand.
Reading back, i actually agree with you on most points. It just seems that you're searching for some sort of 'fairness' in a 'system' where it doesn't really exist.
Just would like to add a distinction that EMH seems to hold from a retail clients perspective, but does not hold from a systematic / structural side.
Markets are structurally exploitable. For example it is (generally) illegal for an institution to front run clients' orders, however it is ok if you are just a trader and know probable (not necessarily sure fire) locations of large orders and front run them.
> It just seems that you're searching for some sort of 'fairness' in a 'system' where it doesn't really exist.
But "fairness" is what the EMH means. "Fairness" expresses in human terms what the mathematics tells us. An efficient market ruthlessly adjusts for inefficiencies and oversights, sort of like biological evolution, and evolution is the ultimate in ruthlessness and indifference to the fate of individuals. It's also fair -- each species has an equal chance to prevail.
Insider trading is unfair to ordinary investors who aren't criminals. The announcement effect is unfair to ordinary investors who don't have that kind of influence over the markets. If you take away insider trading and the announcement effect, the result is ... more fair. That result is closer to an optimally efficient market.
> Markets are structurally exploitable.
Only by cheating (and as your example demonstrates). The SEC knows this, and they know how corrosive distrust in the fairness of equities can be for society as a whole, so they aggressively pursue cheaters.
Absent cheating, equities markets cannot be gamed. Absent cheating, a long-term buy & hold position is the optimal strategy, By the way, and for what it's worth, Warren Buffett agrees.
Markets have many inefficiencies of varying persistence that don't depend on insider knowledge, speed advantages, or cheating, but using simple widely available public information, some on just the price itself.
I think it doesn't validate EMH if such inefficiencies disappear if exposed, if they depend on skill and systemic knowledge rather than 'cheating' - i suppose it depends on where one draws the line between cheating and skill.
Discovering and exploiting them is a matter of study and work which the public is unwilling to do, and actually actively discouraged by the status quo of this industry.
They're guided into ignorance and defering their finances to the sucky professionals who get paid for amount of assets under management rather than performance.
Warren Buffet doesn't buy and hold indiscriminately. He buys things on sale during market crashes when the panicky public gladly sell to him. For example by selling put options he gets paid just for waiting for the discounts.
And as a potentially controvertial aside, i think his size and connections within the government grant him hypocritical priviledges not available to the public (other than through owning his stock of course).
> Discovering and exploiting them [i.e. inefficiencies] is a matter of study and work which the public is unwilling to do, and actually actively discouraged by the status quo of this industry.
Yes, but it's well-established that the average investor comes out behind by trying to exploit such inefficiencies, and this outcome supports the idea that the EMH is true, and that a buy & hold position produces the best returns on average, overall.
The problem with Wall Street anecdotes is that successes are given undue weight in the telling, misleading people into the idea that locating and exploiting inefficiencies puts you ahead. On average, this is false -- very, very false.
> Warren Buffet doesn't buy and hold indiscriminately.
Berkshire Hathaway's long-term success is easily explained -- its position is a matter of public knowledge, so others, hoping to see the same growth, buy the same stocks (too late to benefit themselves but just in time to benefit Berkshire Hathaway). It's called the "announcement effect".
I throw 100 dice 10 times and 2 of them come up 6 every time, i pick them up and call them survivors. Then i throw these 2 survivors 10 times and one of them again comes up 6 everytime - to me its worth the effort to figure out why, regardless if it's branded madoff or rentec or buffett, or that the other 99 exhibit traits of EMH.
Excusing it away feels like a lazy way out, settling for mediocrity. I make a living by throwing sixes time and time again. It's not cheating, it's not easy, it takes skill and work and is doable.
"If you can trade faster than others, you can drain some resources from them"
What's embarrassing is how many intellectual resources are focused on this problem. For those not in finance, compare it to focusing talent on finding better advertising techniques.
Models are approximations. Don't confuse your models with reality.
Know what your assumptions are, and have a plan for what to do when those assumptions no longer hold.
Know why your models work, so that you know when they are likely to stop working.
Get deep domain knowledge.
Know that the market can stay irrational longer than you can stay solvent.
Experiment ruthlessly. Apply the scientific method.
Start simple and iterate.
This probably applies to more than just finance.