More on TA. Does It Work?


Readers’ Replies to prior post on Technical Analysis (“TA”) found here:

#1 Here is one rich technician. Paul Tudor Jones. Nuff said. Lowery research has been in business a long time doing TA. Tom Demark. Look him up.

#2 Personally I think people should use Fundamentals for investing in anything for the Long Term. Technical Analysis has a purpose but usually only for the immediate future. That’s why most Day Traders use Technical Analysis.

For instance, if you watch the Moving Averages and say that the 200-day Moving Average (200MA) falls below a 50-day moving average (50MA), this is known as a death cross and 9 times out of 10 that I have seen that kind of action, the price on that chart will start to go down.

Technical Analysis is mostly used for short-term movement in a stock, commodity, currency, etc., etc. It is virtually impossible to base a long-term investment on Technical Analysis.

My response: Thanks, but those opinions don’t improve our knowledge about whether TA is a usable tool.  Take #1, Paul Tudor Jones is a big, successful hedgie who uses TA–enough said.  Let’s substitute TA for dresses in drag and flips coins. The meaning would be the same.  I don’t want to pick on anyone, I read the same in many articles on Tudor Jones. I bet you Tudor Jones couldn’t even tell you EXACTLY how he uses charts.  He probably blends many factors into his “sixth-sense” based on thousands of hours of intensive interaction with the markets.See page one: 04_Jul_-_Tudor_Inv_Corp where Tudor loves the dollar and then the next day he is short the dollar. Did a chart give him a signal? If so, what is the STATISTICAL EVIDENCE?

The point is, there is and can never be any statistical evidence since charts just reflect PAST human choices of buying and selling. Future human action can’t be mathematically proscribed.

Paul Tudor Jones II Interview

Also, technical analysis has both passionate critics and ardent adherents. For example, an October 2009 study by New Zealand’s Massey University found that of more than 5,000 strategies that employ technical analysis, none produced returns in the 49 countries where researchers tested the strategies beyond what you’d expect by chance. However, scores of traders, including billionaire Paul Tudor Jones, say the discipline helped them amass great fortunes. So I tried to keep an open mind. (If Paul Tudor Jones is a billionaire, then TA must work! Flawed logic!) Read more at

#2 If you have evidence that 9 out of 10 times the “Death Cross” moves prices enough for you to take advantage of them–then great for you. But again, if this “signal” did work, why wouldn’t the market DISCOUNT it in the future especially if you could precisely define what a Death Cross is?

There are some money managers who use TA in creative ways for the long-term but I call them market mystics. 

Capital Mkt Update 2008 Montgomery and Capital Mkt Update 2009 Montgomery

What I AM saying is to use TA if it works for you however you define “works for you” be it in confidence, money management and setting risk parameters, finding opportunities, etc. but don’t fool yourself. THERE IS NO SCIENTIFIC EVIDENCE THAT TA HAS ANY EFFICACY.

I will make a $1,000 bet. Show me any statistical proof or long-term (fifiteen years or more) of market beating returns solely using TA. Ask these guys:  I guess SELLING TA is more profitable than USING it. I smell a legal high-pressure selling scam:

Why the strange picture at the top of this post?   This NY Times’ article by Gary Taubes shows how difficult it is to obtain scientific proof for even life threatening health issues.

NEARLY six weeks into the 2014 diet season, it’s a good bet that many of us who made New Year’s resolutions to lose weight have already peaked. If clinical trials are any indication, we’ve lost much of the weight we can expect to lose. In a year or two we’ll be back within half a dozen pounds of where we are today.
The question is why. Is this a failure of willpower or of technique? Was our chosen dietary intervention — whether from the latest best-selling diet book or merely a concerted attempt to eat less and exercise more — doomed to failure? Considering that obesity and its related diseases — most notably,Type 2 diabetes — now cost the health care system more than $1 billion per day, it’s not hyperbolic to suggest that the health of the nation may depend on which is the correct answer.
Since the 1960s, nutrition science has been dominated by two conflicting observations. One is that we know how to eat healthy and maintain a healthy weight. The other is that the rapidly increasing rates of obesity and diabetes suggest that something about the conventional thinking is simply wrong.
In 1960, fewer than 13 percent of Americans were obese, and diabetes had been diagnosed in 1 percent. Today, the percentage of obese Americans has almost tripled; the percentage of Americans with diabetes has increased seven-fold.
Meanwhile, the research literature on obesity has also ballooned. In 1960, fewer than 1,100 articles were published on obesity or diabetes in the indexed medical literature. Last year it was more than 44,000. In total, over 600,000 articles have been published purporting to convey some meaningful information on these conditions.
It would be nice to think that this deluge of research has brought clarity to the issue. The trend data argue otherwise. If we understand these disorders so well, why have we failed so miserably to prevent them? The conventional explanation is that this is the manifestation of an unfortunate reality: Type 2 diabetes is caused or exacerbated by obesity, and obesity is a complex, intractable disorder. The more we learn, the more we need to know.
Here’s another possibility: The 600,000 articles — along with several tens of thousands of diet books — are the noise generated by a dysfunctional research establishment. Because the nutrition research community has failed to establish reliable, unambiguous knowledge about the environmental triggers of obesity and diabetes, it has opened the door to a diversity of opinions on the subject, of hypotheses about cause, cure and prevention, many of which cannot be refuted by the existing evidence. Everyone has a theory. The evidence doesn’t exist to say unequivocally who’s wrong.
The situation is understandable; it’s a learning experience in the limits of science. The protocol of science is the process of hypothesis and test. This three-word phrase, though, does not do it justice. The philosopher Karl Popper did when he described “the method of science as the method of bold conjectures and ingenious and severe attempts to refute them.”
In nutrition, the hypotheses are speculations about what foods or dietary patterns help or hinder our pursuit of a long and healthy life. The ingenious and severe attempts to refute the hypotheses are the experimental tests — the clinical trials and, to be specific, randomized controlled trials. Because the hypotheses are ultimately about what happens to us over decades, meaningful trials are prohibitively expensive and exceedingly difficult. It means convincing thousands of people to change what they eat for years to decades. Eventually enough heart attacks, cancers and deaths have to happen among the subjects so it can be established whether the dietary intervention was beneficial or detrimental.
And before any of this can even be attempted, someone’s got to pay for it. Since no pharmaceutical company stands to benefit, prospective sources are limited, particularly when we insist the answers are already known. Without such trials, though, we’re only guessing whether we know the truth.
Back in the 1960s, when researchers first took seriously the idea thatdietary fat caused heart disease, they acknowledged that such trials were necessary and studied the feasibility for years. Eventually the leadership at the National Institutes of Health concluded that the trials would be too expensive — perhaps a billion dollars — and might get the wrong answer anyway. They might botch the study and never know it. They certainly couldn’t afford to do two such studies, even though replication is a core principle of the scientific method. Since then, advice to restrict fat or avoid saturated fat has been based on suppositions about what would have happened had such trials been done, not on the studies themselves.
Nutritionists have adjusted to this reality by accepting a lower standard of evidence on what they’ll believe to be true. They do experiments with laboratory animals, for instance, following them for the better part of the animal’s lifetime — a year or two in rodents, say — and assume or at least hope that the results apply to humans. And maybe they do, but we can’t know for sure without doing the human experiments.
They do experiments on humans — the species of interest — for days or weeks or even a year or two and then assume that the results apply to decades. And maybe they do, but we can’t know for sure. That’s a hypothesis, and it must be tested.
And they do what are called observational studies, observing populations for decades, documenting what people eat and what illnesses beset them, and then assume that the associations they observe between diet and disease are indeed causal — that if people who eat copious vegetables, for instance, live longer than those who don’t, it’s the vegetables that cause the effect of a longer life. And maybe they do, but there’s no way to know without experimental trials to test that hypothesis.
The associations that emerge from these studies used to be known as “hypothesis-generating data,” based on the fact that an association tells us only that two things changed together in time, not that one caused the other. So associations generate hypotheses of causality that then have to be tested. But this hypothesis-generating caveat has been dropped over the years as researchers studying nutrition have decided that this is the best they can do.
One lesson of science, though, is that if the best you can do isn’t good enough to establish reliable knowledge, first acknowledge it — relentless honesty about what can and cannot be extrapolated from data is another core principle of science — and then do more, or do something else. As it is, we have a field of sort-of-science in which hypotheses are treated as facts because they’re too hard or expensive to test, and there are so many hypotheses that what journalists like to call “leading authorities” disagree with one another daily.

It’s an unacceptable situation. Obesity and diabetes are epidemic, and yet the only relevant fact on which relatively unambiguous data exist to support a consensus is that most of us are surely eating too much of something. (My vote is sugars and refined grains; we all have our biases.) Making meaningful inroads against obesity and diabetes on a population level requires that we know how to treat and prevent it on an individual level. We’re going to have to stop believing we know the answer, and challenge ourselves to come up with trials that do a better job of testing our beliefs.

Before I, for one, make another dietary resolution, I’d like to know that what I believe I know about a healthy diet is really so. Is that too much to ask?
Gary Taubes is a health and science journalist and co-founder of the Nutrition Science Initiative.

16 responses to “More on TA. Does It Work?

  1. Markets are constantly changing. A single technical indicator with fixed parameters cannot work indefinitely. That’s what academicians test and find to be not working. That much is fairly well-established.

    On the other hand, if you have an ensemble of technical indicators, if you have statistical prowess to distinguish working from not working, if you adapt those technical indicators to the current market environment & asset class you’re trading, I believe you can gain an edge with technical indicators.

    I believe TA is a pre-cursor to quant trading (= time series models, machine learning models etc). Quant trading does work… there are technical/time-series patterns in financial time-series. There are well-documented hedge funds & prop trading firms who consistently make money doing quant trading. Can an individual gain the same edge? Probably not. It requires significant expertise and know-how to make such models work. It’s also a full-time job, you can’t quite pursue this as a hobby successfully. At best you’ll give up trying after some loss. At worst, you’ll lose years trying without any tangible results. An individual can’t easily replicate institutional know-how (= a decade of research by 20 researchers = 200 man-years for an individual to catch up).

    If quant trading works, can TA be made to work? I think the answer is yes. Like I said, the ability to distinguish the wheat from the chaff is key. The way you optimize/adapt your models & parameters is also extremely important. In fact, I know enough people who’re doing it successfully (again in an institutional environment, diversified).

    So, yes, TA can be made to work. But not in the way academicians or lay people imagine.


  2. On a separate, note I’ve read Gary Taubes’ 600-page tome (Good Calories, Bad Calories). I think the guy deserves a PhD for that (at least, an honorary one). It’s a remarkable study of the state-of-the-art in nutrition. His writings and analyses are highly recommended.

    If you’re more into the business of processed food, you may want to start with Salt Sugar Fat: How the Food Giants Hooked Us by M. Moss—another book I highly recommend. It has excellent investment insights.

    Reading these books might be difficult (they’re long), I personally recommend listening to them on Audible, a smart-phone app. Audible changed my life, hopefully it’ll change yours as well (listened to 10+ books in the last few months). The best way to gain worldly-wisdom, as Munger would put it.

  3. John,

    What do you mean by “future human action can’t be mathematically proscribed”?

    Your arbitrage argument (“why wouldn’t the market DISCOUNT it in the future”) could just as easily be applied to Fundamental Analysis (FA), where it appears to be wrong. So why should it be valid for TA?

    FA exploits human psychology, that is the main reason why it doesn’t get arbitraged away. Why do you categorically rule out that human psychology can be exploited by TA?

  4. Gentlemen, let me offer an alternative explanation as to why TA and/or value investing works.

    The most fundamental function of financial markets is liquidity provision. Namely, on any given *day*, the number of buyers won’t exactly match number of sellers, then you need an intermediary to come in and absorb the excess shares. Similarly, on any given *week*, (buyers != sellers) => an intermediary has to come in and absorb the excess. Similarly, on any given *year*, (buyers != sellers) => an intermediary has to come in and absorb the excess.

    I guess you get the idea: in any time frame, intermediaries are needed to “provide
    liquidity” to the market.

    Intraday: intermediary = HFTs.
    Multi-day: intermediary = stat-arb hedge funds, e.g. Medallion fund by Renaissance Technologies.
    Multi-year: intermediary = value investors like Warren Buffet, who provided liquidity to the market in the darkest days of ’08-’09 by buying bank and other shares.

    A pattern must be getting apparent to you: All these intermediaries need to buy low and sell high to stay in business, i.e. they need to ascertain “value” in their time-frame, buy when there’s “value” and put enough “margin of safety” before getting in or out.

    You might say: I understand what value means for likes of Warrent Buffet, but *what is value for HFTs and multi-day stat arb strategies?* Well, that value is usually quantified through time series analysis and modeling. And, yes, don’t be skeptic, this is _the_ fundamental liquidity provision mechanism in the markets. One way or another [some] hedge funds, prop trading firms, HFTs bet on mean reversion in time or across instruments to provide liquidity and make profits.

    Without such mean-reversionary bets, prices would oscillate wildly and markets would be extremely inefficient. Sort of like what’s happening to bitcoin, as bitcoin cannot be hedged, hence the liquidity provision is a lot more difficult in bitcoin than in equities. Also, sort of like what happened during 2010 Flash Crash, where liquidity providers pulled out of the market and prices oscillated wildly between zero and infinity.

    Now, what’s the relation to TA? Well, TA is just a precursor to quant trading. You can find a ton of technical indicators to quantify the degree of price oscillations & “margin of safety” (i.e. deviation from some value estimate). Before time-series methods, people did use TA to quantify oscillations. Certain indicators did work for extended periods (though ultimately more modern time-series prevailed, as they are a lot more capable). Hope you’re getting the picture.

    I’ve written about oscillatory technical indicators so far, there are also trend indicator. Yes, trend indicators do also work. Why? Because good news/bad news about the economy are clustered, i.e. because the business cycle tends to be multi-year long, good news tend to be followed more good news, and vice versa. The same applies to individual companies.

    Market, when left by itself, doesn’t adjust fast enough to turning points (remember what happened in 2009 after the crisis turned). So *huge* trends happen thanks to continuity of business cycle, and news clustering. By just watching the SPY chart you can see this yourselves. Now, that doesn’t mean profit is possible or easy. Only the most sophisticated, experienced trend followers can get such systems to work and eventually prevail. It’s all a Darwinian process: Bad trend followers get eliminated, and the ones who survived tend to be the best ones (not just the lucky ones). Does this make sense?

    I know enough people who run such strategies successfully (though they rarely make the news like Buffett). The quant/time-series versions should be preferable to TA, but TA, when done right, can be made to work. There are fundamental reasons why TA should work.

    Sorry for the long post. Hope this clarifies things somewhat.

  5. If words alone are not enough, and you’re still skeptic (I personally would be!), here’s a bit of data. CTAs run trend following strategies and most of them use TA. If Barclay’s record is to be believed, these strategies have had a pretty good run:

    I personally wouldn’t put my money into any of these 2/20 hedge funds due to excessive fees. But, I fundamentally believe such trend and/or liquidity provision strategies can be made to work with the right kind of people running them.

    • I think Richard Pzena in one of Prof. Greenblatt’s classes said momentum investing works. Studies confirm. A trend persists especially when markets have government intervention like currency markets. Pzena said that momentum investing was too difficult for him to use. Emotionally, you either love bargains or following trends. Perhaps there are those who can combine both. Thanks for the thoughtful post.

  6. Discretionary TA does not work as it is too subjective. Mechanical TA might work if it has been complemented with rigorous back-testing, optimization and forward-walk analysis i.e. followed the scientific method!

    For example, a classic momentum trade is where the 20-day MA crosses the 50-day MA. Does anyone know why 20-day and 50-day have been used? Does anyone bother to ask that question? Why isn’t 21-day (a trading month) or 84-day (a trading quarter) MA used? Simply, because in the era of no computers when things needed to be done by hand, it was easier to have whole numbers as MA.

    These MA rules/parameters have since been passed down to retail traders who do not go through rigorous scientific method of testing the rules. And then they find out it doesn’t work for them now.

    Currently, no hedge fund uses this momentum strategy anymore. Instead they use the more academic version of momentum where portfolios are formed and stocks are sorted into deciles.

  7. Mohammed Al-Alwan

    Hi John
    An excellent post on a very controversial and debated subject. You have highlighted a couple of valid points and I want to share with my thoughts some articles I collected on the subject. Although, many people attack TA and claim such approach does not help generating Alpha as there are not many successful people practicing it, I believe same things can be applied to FA. More than 90%of money managers and novice investor underperform the markets depending on which study you read. So, to be objective I think investors should focus on what’s working for them rather than what others claim is working.
    First I would like to provide links to selective articles written by leading academic on the topic. Interestingly, father of efficient market hypothesis is now acknowledging that momentum is a significant factor in explaining equity returns and updated his seminal work Fama and French three factor model into a four factor model. Also, I would recommend watching his interview in CFA equity conference held I think in 2012 where he said “the most embarrassing development since EMH started is not behavioral finance but the persistence of price momentum”.
    The following are articles discussing the momentum anomaly.

    Track Record of Some famous traders and Trend follower. Some of them are billionaire and has track record that spans decades

    Famous Money Managers who employ both TA & FA.
    1. Michael Burry the famed value investor who shorted subprime is also using TA. See this great post by Street capitalist
    2. Although Stanley Druckenmiller is macro trader he has mentioned many times his reliance on TA. “
    3. Legendary fidelity money manager Anthony Bolton and author of two books mentioned many times his reliance on TA for position sizing and portfolio construction

    4. John palicka the author of fusion investing book has also long track record using both TA & FA and authored a book on the subject .Although I read the book and I did not find it well written
    5. Martin Pring the famed Technician has also a good track record for more than a decade.
    6. Charles Kirkpatrick has been using both TA & FA successfully

    Greenwald wrote in his book value investing that there are many strategies that outperform the market including TA. However, for average investors and professionals as well, value investing, is the most efficient and reliable method. So, the discussion should be focused on which method is more reliable and efficient rather than which method works.

  8. An award winning comment. This probably needs to be reposted so more are aware of it. As soon as I take time off from buying precious metals stocks, I will read all those articles. Thanks.

    At the end of the day, you have to chose a method(s) that YOU are comfortable with and can persevere through adversity while knowing what will give you an edge.

  9. Mohammed Al-Alwan, read my previous comment please. It addresses exactly what you are talking about. There is a difference between looking for a ‘bearish engulfing pattern hammer’ and all that non-sense used in a discretionary way versus using bunch of indicators with excellent money management rules and exit rules in a systematic way with no emotions i.e. using a computer.

    Again, the momentum that is embarrassing to EMH is COMPLETELY DIFFERENT to the momentum that technical analysts talk about (see my example on moving averages).

  10. Mohammed Al Alwan

    Dear Terminator

    The debate is on terminologies i would assume.Technically its defined as serial correlation in the time series.if you put on statistical lens you will call it serial correlation.But through Technical lens you will call it trend .Same is true if you use Academic terms you will call it momentum.That is why EMH theory states in the weak form if market is efficient then technical analysis does not work.Fama was not talking about trend lines,MA, etc.He was just focusing on the serial correlation in times series.Either using your naked eyes or preferably using more robust statical methods.but conclusion is the same.

    i agree that discretionary methods are pron to biases and subjective interpretation but the objective is the same in both discretionary or mechanical.

  11. Mohammed Al Alwan

    Here is a summary link to Fama interview in the CFA Annual Conference
    where “said Fama. “What is a momentum stock today probably isn’t one three months from now. Of all the things that I think are potential embarrassments to market efficiency that is the primary one. It’s just a model,” he said.”

  12. Mohammed Al Alwan

    valid comment on Fama EMH from Margin of Safety blog by Ray Galkowski, CFA:

    One has to admire Fama’s consistency. He is wrong about market efficiency, but he is consistent.

    Although he makes some good points, he also makes many errors. His criticism of Kahneman is absolutely wrong, for example. Kahneman would say that we are both rational and irrational at times too.

    It is most interesting to me that Fama (and French) could write a great paper that exposed several significant flaws in the efficient market hypothesis; accept significant well-paid positions at an investment firm–DFA–that exploited the inefficiencies that F&F exposed and turned DFA into a juggernaut after it had floundered for years; and that Fama could still stand before 1,800+ delegates at the CFA Institute Annual Conference and say with a straight face the markets are efficient.

    Sebastian Mallaby also spoke at the conference and he talked about how Paul Samuelson would defend market efficiency until his last breath, but that when it came time to invest his own money he placed it with Commodities Corp., a Princeton, NJ firm that attempted to exploit market inefficiencies. Mallaby’s conclusion: “Invest as they invest and not as they say to invest.” Good advice. Lecturing that markets are efficient and that we are fools to try to find bargains is a good way to hoard the limited number of great opportunities for oneself.

  13. Dear Mohammed Al Alwan, thank you for your replies. I love this discussion. My discussion below of momentum refers to the academic version.

    Excellent point and here is the crux of my argument. Serial correlation would have been exploitable before the digital/information age through discretionary trading using technical analysis. But nowadays you cannot extract that serial correlation into profitable trading strategies by just using the eye.

    There is empirical evidence of positive and negative return autocorrelations(i.e. the correlation of last period’s return with this period’s) at different time horizons in different asset classes and various countries. However, this alone DOES NOT lead to superior predictability through the use of trading strategies. The superior predictability (i.e. momentum anomaly) comes from forming portfolios in a certain, mechanical way. Here is a youtube video of John Cochrane explaining momentum on MINUTE 47. His analogy of ‘buckets’ is excellent.

    Autocorrelation/Serial correlation is just a necessary precondition for momentum to exist but it is not a conclusive proof of the anomaly of momentum.

  14. Kirtzpatrick book has a link to a comprehensive paper on TA. 50 something techniques worked hustorically while 20 something didn’t – everything out-of-sample, of course.

    Momentum clearly works and it’s actually more powerful than P/B and the small cap factor, according to Carhart’s research. Short term momentum is even a fact for classic value investors such as Haugen and Greenblatt – earnings surprises are a well studied market anomaly with a bunch of studies that appear even in the CFA curriculum.

    TA can work. Jim Simmons has the best track record of the last 20 years or so ( about 70% gross per year) and is much more into TA than value.

    I like value – I think it’s less vulnerable to copycats and self fulfilling prophecies, but there are many ways to make money and I respect them all

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.