Technical Analysis ... Astrology?

The basic thing to keep in mind is that there are many different ways to make money trading. However, buying trading signals is probably the worst thing anyone can do.

Instead, it is much better to formulate a trading strategy that works for you. Some TA guys do TA very well but can't algo trade. Some people it is just the opposite. Even Soros claims he physically feels it when something is brewing in the markets and he acts accordingly. So honestly who is going to say that he is wrong?

But again, stay away from the snake oil salesmen!

Considering Market Wizards was about traders who made hundreds (and billions in some cases) of dollars trading various styles answers you own question. There is no shortage of ways to make a fortune trading. Some only use technical analysis while some are only fundamentalists. The main point (and one that was covered a lot in both books) is that you match your personality and find the style that is right for you.

If you think about it from a psychological point of view, some of it makes sense. Just for example, resistance levels. They exist because people have a point where they feel that they have made their money and are willing to exit the market. 10k for the DOW was one. It's a nice big round number.

By no means am I saying that you should go out and start trading on every signal you read, but you also shouldn't just dismiss it completely

In my opinion, it is at least to some extent a self-fulfilling prophecy. I know lots of people who agree with that idea, including those who trade at least partially using TA.

http://i43.tinypic.com/2cy4rb6.jpg Guess there's only one way to settle this... I've signed up for a spread betting account and will toy around with support/resistance lines (starting with gold vs usd as above). If this time next year I'm a millionaire we'll assume technical analysis has value.

Everyone has a preference on what info to look at, but I don't know any good traders who ignore technicals altogether.

People adverse to technicals need to consider that every 'fundamental' model is built on a set of assumptions (revenue growth, margins, etc.).

There is an arbitrary factor no matter what approach you take - if there isn't, then you've found pure arb.

http://i39.tinypic.com/os5wk1.jpg Looks like a few others had similar ideas. Though funnily enough I still lost money on this - too much adrenalin in the blood...

http://i41.tinypic.com/2e496qh.jpg

Different people see different things....

Technical analysis is bullshit. Period. Full stop.

People sell (and buy!) "systems" for picking lottery numbers too. Complete and total bullshit from start to finish. And just because a lot of people engage in bullshit, doesn't mean you have to pay any attention. The idiots who sit in their underwear and "do some analytics," as the annoying, hipster-voiced E-Trade baby says, are not a large enough collective force to move the markets.

I've closed my spread betting account already. I believe any attempt to consistently 'win' by placing bets based purely on support/resistance breakouts is plainly a fools errand. But I also agree with JSin that to ignore such 'chart developments', for lack of a better name, is a mistake.

http://www.amazon.com/Random-Walk-Down-Wall-Street/dp/0393315290

Suggests strongly that technical analysis is useless.

A Random Walk Down Wallstreet is about investing, not trading. If you're looking to buy and hold for long periods of time, then right, technical analysis is useless. But day trading is a completely different game, and it definitely helps. The big prop firms use technical analysis as one of their tools, and that's enough to convince me

time frames are important- surprised nobody mentioned products too. I think....Even if one doesn't use it directly TA is something you should consider with SHORT-term trading in highly-liquid, robust markets. The most common example is FX.

FX is definitely the most likely to trade technicals.

why is FX the most likely to trade technical? It seems it would be much more fundamental based (considering the role of central banks, international capital flows, etc.)

Just due to sheer liquidity and short-term bias.

From my understanding, most evidence and research suggests it is fantasy. But there is evidence of cases where, likely due to a self-fulfilling prophecy effect, it is a successful method. Such an uncertain effect is probably not a good idea to use as the basis for your investing, however.

Just my two cents.

A small firm I interned at used 99% technical analysis and beat the market since inception. I know for a fact it can work. Will the average joe be successful with it? No. The average joe can't even beat the market using fundamentals.

if it didn't work to some extent then it probably wouldn't exist. i think there is some truth behind most anything, its how that truth is utilized that lies the value.

Technical analysis in my opinion by itself would not be so accurate, market is like a big puzzle and technical analysis in many of the cases only follows the market "After" everything passed, so used alone i dont think would give you the results you expect, also it is very difficult to forecast the impact in the market of an economic data as one person mentioned the average Joe will not make it, its really hard and takes time.

No, tech analysis is very useful, especially during downturns

50% of the time, it works every time.

Technical Analysis works until it doesn't. I do really look at technicals quite a bit. HOWEVER -- never underestimate how much one Portfolio Manager can dislocate a stock. With this in mind (and as you've probably read) the longer the time horizon the more valid the signal.

If your talking about scientific T.A. in which you create a hypothesis and test it empirically, then yes, it works. In that case a lot of quants would (technically) be technical analysts as well.

However if you think simply looking at a chart, and buying based on support and resistance, or trend following (just buying a stock because its been up a lot in the past few weeks), then its pure silliness. No succesful hedge fund manager, bank or prop shop, that ive heard of makes trade decisions based on this form of T.A. involving MACD bullshit. In fact there is almost no empirical evidence (that i know of at least) in favor of this form of T.A. All the evidence i've seen is purely anecdotal.

Professional prop trader here. Technical analysis works. FACT. That doesn't mean it is easy or obvious. Most empirical tests are too simplistic. Of course there is not edge simply buying at any kind of support. That is like saying fundamental analysis is bullshit because you tried simply buying stocks with X ratio > 1. Can a person read a few basic books about fundamentals and beat the market? No..... So a person having read a few books about technicals can't beat the market. It takes years of deep thought and understanding to develop these skills to a point where they have statistically proven edge.

So many people in I-Banking or WSO keyboard specialists will say it is like astrology. My brother is in banking and all of his co-workers think my job is a random walk. They're just ignorant. If anybody wants to say otherwise, put your money where your mouth is and let's structure a bet.

It works, because its a self fulfilling prophecy

Its not going anywhere

Please read my previous comment. Can you find edge in fundamental analysis by simply buying companies that have X ratio of a certain value? No, yet I'm not stupid enough to tell Benjamin Graham or Warren Buffet that their method is bullshit, simply because studies show simply buying stocks with a certain asset:liability ratio doesn't work. You need a much deeper understanding and more advanced criteria. Once again, of course a simple trendline is too simple. Most empirical tests that I have read do not apply advanced criteria and display an understanding of the methods.

Of course certain quant shops aren't interested in chart patterns. At the same time, it shows your ignorance, since there are certain very reputable market making firms that value that knowledge, particularly in this lower-volatility environment.

As a simple example, it is hard to argue that Paul Tudor Jones has been quite successful. In his documentary, there is a great short trade he put on due to the chart pattern being similar to that of the Great Depression crash.

Anywho, shoot me a PM and we'll discuss various criteria that you would count as hard evidence.

Also i pm'ed you the question: what would it take to disprove the T.A. works? Can T.A. be fasifiable. can you post the response here.

Technical analysis is complete BS. It works 100% of the time in hindsight ("Of course I made a bad trade -- how could I have missed that clear head-and-shoulders signal?!? It's so obvious now! The chart is never wrong.").

shit guys thanks gonna tell all our traders and pms on tuesday that they should scrap all of the charts and graphs and rely completely on fundamentals because that's exactly what the market is valued at right now, right? guaranteed to get a promotion after that, i'll hire one of you guys as an intern for bestowing this knowledge upon me

whoa, thanks again dude. i'll be sure to tell everyone that we're no longer a reputable shop and the fuckton of money we made this year really doesn't exist because we don't follow some shit a college kid says on a message board. where would i be without you bro?

Once again, this is exactly what im talking about. Technical analysts generally have no evidence to back their claims other than telling you how much money they made in a bull market. All completely anecdotal.

And again I ask, what would it take to falsify technical analysis?

completely dismissing technicals based on some tea leaf voodoo shit is extremely short-sighted i never said we rely solely on technicals. i'm just saying that if you completely rely on fundamentals (especially in a market like this), you are going to going to get pounded into the ground. the same thing goes for solely relying on technicals

I don't get why you assume I don't test my strategies for statistical edge? All of what I do is statistically proven to work or I wouldn't use it lol. I do the same exact thing Paul Tudor Jones does and what you promote so much. I form a hypothesis and I test it historically and in real-time. These strategies using technical analysis can have edge and do have edge or else I would not have a job, nor would my co-workers.

Working on a PM response now.

once again, what would it take to falsify the type of t.a. you speak of?

Does anyone else see a fundamental problem with backtesting in the context of the market? It's not a controlled scientific experiment where you can exactly recreate the conditions of your experiment -- the market is always changing.

I agree that patterns frequently repeat themselves...but they also frequently don't. And the circumstances of each are always different.

I provided to jackbnimble, through PM, the name of a major oil firm I worked for that some of their paper trades employed TA. I also provided Richard Dennis as another hedge fund who used TA and also the name of a market making firm that finds TA to be useful. Also, I provided him the name of a BB trader that uses TA.

I also provided him a link to an empirical study available to the public that helps me build intuition about certain things. Other details regarding specifics about my background and what I've experienced I would prefer to keep private.

Everybody knows about the success of Richard Dennis and the Turtle Traders...

"Another of the Turtles, Michael Carr, a former fantasy game designer found Dennis' ad in a newspaper the day he'd lost his job. Michael had no trading experience at all, but shortly before getting fired he conceived an idea to create a game based on stock trading. He'd read several brochures and visited 6-day evening courses, so he did have a rudimentary idea of what market is. He was probably one of the few applicants who didn't know who Richard Dennis was - a funny accident happened when Carr asked Dennis whether he used technical or fundamental analysis. Dennis couldn't help giggling - everybody knew he used strictly technical analysis." Modern Trading Magazine

Jackbnimble made the claim that BBs don't use technical analysis. As further support to how absolutely stupid this claim is, let me provide the following link which references David Sneddon who I referred him to in PM. I'm not sure how you can say bulge brackets don't believe in technical analysis, when right here is a list of TA guys from everything from UBS to RBC to CS. Are these not real banks? Lmfao.

"TECHNICAL ANALYST OF THE YEAR Richard Adcock (UBS Investment Bank) Peter Beuttell (MTS Research) George Davis (RBC Capital Markets) Paul Desmond (Lowry Research) Jean-Charles Gand (BBSP) Valerie Gastaldy (DayByDay) Remy Gaussens (Trading Central) Andrew Gebhardt (Finex) Peter Hafez (RavenPack) Karen Jones (Commerzbank) Martin Jones (Informa Global Markets Steve Miley (Credit Suisse) Riccardo Ronco (Aviate Global) David Sneddon (Credit Suisse) Ron William (MIG Bank) WINNER: Riccardo Ronco (Aviate Global)"

http://technicalanalyst.co.uk/conferences/Awards12_finalists.htm

Here is a research piece from RBC discussing technical support and resistance levels, trendlines, and moving averages in FX. Again, it is so absolutely asinine to say BBs don't use TA.

http://technicalanalyst.co.uk/Research/RBC-15May2013.pdf

sorry i posted the above before reading the whole thread. Let me put one debate to rest: Traders at banks do indeed use very simple technical analysis all the time. They also do hire technical analysts, some of whom are very well followed and respected by smart, sophisticated, buyside clients and some of whom are very well-followed internally by market-making desks.

One can say what they want about that, but it is the truth.

Every investor should use their fair share of fundamental AND technical analysis. They both have their purpose and proper times of use. I would not make a trade that I am going to make and hold for 6 hours based off of some shmuck's DCF model or how much a company's top line has grown over 5 years.. Proper and smart technical analysis is meant for these types of trades. They both have their time and place. A good investor knows how/when to use both.

jackbnimbleOnce again. Goldman Sachs traders are not building algorithms based on moving averages. neither is rentech, or Getco, or Jane Street, or ctc or Citadel.

Many PMs at those above HFs use momentum indicators extensively. I'm saying that not because I've spoken to an adequate sample of PMs at the above funds, much less one that would disclose their methods extensively, but because such indicators are rather popular among systematic macro traders. Kaveh Alamouti, Citadel's head of macro trading, once hired a bunch of trend following traders to manage money in the mid 2000s while at Moore Europe, including a famous 'turtle'. Rentech's Medallion fund also purely focused on TF strategies until the early 90s, at which point they transitioned into short term trading. Every single large macro fund I can think of, from old school Caxton to the more modern Brevan Howard, run an exclusive TF fund as well, and most of the time their entry signals involve using methods that are very similar to moving average crossovers, if not the same thing. Hell, even pure quant shops like TwoSigma and WorldQuant have managers running such strategies (publicly disclosed in the case of TwoSigma). I hope it is clear that you are thoroughly misinformed.

Macro Arbitrage jackbnimble:

Once again. Goldman Sachs traders are not building algorithms based on moving averages. neither is rentech, or Getco, or Jane Street, or ctc or Citadel.

Many PMs at those above HFs use momentum indicators extensively. I'm saying that not because I've spoken to an adequate sample of PMs at the above funds, much less one that would disclose their methods extensively, but because such indicators are rather popular among systematic macro traders. Kaveh Alamouti, Citadel's head of macro trading, once hired a bunch of trend following traders to manage money in the mid 2000s while at Moore Europe, including a famous 'turtle'. Rentech's Medallion fund also purely focused on TF strategies until the early 90s, at which point they transitioned into short term trading. Every single large macro fund I can think of, from old school Caxton to the more modern Brevan Howard, run an exclusive TF fund as well, and most of the time their entry signals involve using methods that are very similar to moving average crossovers, if not the same thing. Hell, even pure quant shops like TwoSigma and WorldQuant have managers running such strategies (publicly disclosed in the case of TwoSigma). I hope it is clear that you are thoroughly misinformed.

1) I have also spoken to people at some of these firms albeit not every single one you or I have mentioned. I assure you our conversations did not involve flag patterns or moving averages. In fact (though I am not saying they disclosed to me what their firm was doing) the vibe if you will, and the context and setting in which I was speaking to them indicated that they use the form of TA which I argue in favor of, the one where you test your ideas on empirical evidence to see if there is any basis in reality.

2)im pretty sure I acknowledged this when lbrest posted this. I understand that guys like druckenmiller have also spoken TA. But the vast majority of the profits (as in more than half) made at these firms are not made by blindly following trend lines. In fact with the exception of Goldman Sachs and Citadel, the prop shops I mentioned do not make the bulk of their profits off of blind TA. It seems to me that given the growing demand for tick data for example, and the fact that brokers sell their order flow to these firms, that they care deeply about testing their ideas, otherwise this information would be useless thus yielding no demand.

3)I never contested the history of rentech from the 90's, it seems to me they have chosen to pursue methods TODAY that are not related to momentum indicators. In fact Sebastian Mallaby mentions either rentech or shaw being able to make money exploiting anomalies. Anomalies by definition have to have empirical significance to be called anomalies. Moreover, It seems to me that rentech pursues a largely quantitative strategy given that he is a mathematician that hires other PHds in quantitative fields, keep in mind we had a major bull market during the 90's. DE Shaw has also explicitly mentioned the use of testing.

4) It appears I need to do more hw, with the the macro HF's. On the other hand the fact that most of the prop shops do not rely on TA for the bulk of their profits, is not wrong. But keep in mind I did not mention any macro HF's to begin with. Anyhow, I would probably wager that many of these hedge funds that use the momentum indicators, would likely do just as well, without the momentum indicator, as long as they are trading the direction of the market for whatever period that indicator is adjusted for. The reason I say this is because the markets that macro traders trade (fx, commodities and bonds), have slightly positive autocorrelations.

5) again i am not against the use of technical analysis, but I am against blindly selling a head and shoulders without first quantifying it and testing it.

6) the markets change, if moving average crossovers have worked for a long time in the 90's, they may or may not work now. Actually most of what you post in rebuttle to what my point was were examples from pre 2010. the only exceptions you mention were brevan and 2 sigma. Of course new opportunitities open up, and I have very good gut feeling that these two funds (especially 2 sigma) do not simply make decisions without testing them.

2011 and 2012 were tuff time for the simple momentum guys (of which there are still many)...however, long-term many of these very simple strategies test out very well. I am skeptical of all these algo strategies now, even very complex ones that back-trest impeccably, because there is a ton of money in the space right now and there is too much data that I consider meaningless to current markets (ie testing data from 1992 to derive a strategy to trade interest rates in the new world of central bank interventionism, etc).

But in general I have worked with some people who are considered pretty cutting edge in terms of evidence based-TA-type trading systems and you all would be surprised at how simple some of the stuff they use is. Simplicity is good not bad...as DeVinci said "simplicity is the ultimate sophistication".

Bondarb

2011 and 2012 were tuff time for the simple momentum guys (of which there are still many)...however, long-term many of these very simple strategies test out very well. I am skeptical of all these algo strategies now, even very complex ones that back-trest impeccably, because there is a ton of money in the space right now and there is too much data that I consider meaningless to current markets (ie testing data from 1992 to derive a strategy to trade interest rates in the new world of central bank interventionism, etc).

But in general I have worked with some people who are considered pretty cutting edge in terms of evidence based-TA-type trading systems and you all would be surprised at how simple some of the stuff they use is. Simplicity is good not bad...as DeVinci said "simplicity is the ultimate sophistication".

one thing i've noticed about more simplistic methods is that they tend not to have very high winning percentages. I mean sure they can make money on average, but they are only right more often than not, rather than 90% of the time or more. i personally prefer high winning rates even if I have to give up some performance.

on the other hand if you're not over fitting the data (which is sometimes tough no to do), in my experience, more complexity will yield a higher win rate, but might have slightly lower returns overall.

BondarbSimplicity is good not bad...as DeVinci said "simplicity is the ultimate sophistication".

This is profoundly true, IMO.

jackbnimble Bondarb:

2011 and 2012 were tuff time for the simple momentum guys (of which there are still many)...however, long-term many of these very simple strategies test out very well. I am skeptical of all these algo strategies now, even very complex ones that back-trest impeccably, because there is a ton of money in the space right now and there is too much data that I consider meaningless to current markets (ie testing data from 1992 to derive a strategy to trade interest rates in the new world of central bank interventionism, etc).

But in general I have worked with some people who are considered pretty cutting edge in terms of evidence based-TA-type trading systems and you all would be surprised at how simple some of the stuff they use is. Simplicity is good not bad...as DeVinci said "simplicity is the ultimate sophistication".

one thing i've noticed about more simplistic methods is that they tend not to have very high winning percentages. I mean sure they can make money on average, but they are only right more often than not, rather than 90% of the time or more. i personally prefer high winning rates even if I have to give up some performance.

on the other hand if you're not over fitting the data (which is sometimes tough no to do), in my experience, more complexity will yield a higher win rate, but might have slightly lower returns overall.

yes the old school trend-followers tend to have lower hit ratios but their average winner dwarfs their average loser which is where the edge is derived...this is classic long vol trading. I tend to think that very high win rate strategies are dangerous because they are often short vol, mean reverting strategies that can produce big returns in sample but can blow up in epic fashion when we get the once a decade "crash of a lifetime"...or even just a bad blip on the radar. I similarly am always suspicous of high sharpe ratios. I am a discretionary trader, but I aim for a fairly low hit ratio and when i notice my hit ratio is getting too high I focus on making sure Im being aggresive enough with winning trades.

Bondarb jackbnimble: Bondarb:

2011 and 2012 were tuff time for the simple momentum guys (of which there are still many)...however, long-term many of these very simple strategies test out very well. I am skeptical of all these algo strategies now, even very complex ones that back-trest impeccably, because there is a ton of money in the space right now and there is too much data that I consider meaningless to current markets (ie testing data from 1992 to derive a strategy to trade interest rates in the new world of central bank interventionism, etc).

But in general I have worked with some people who are considered pretty cutting edge in terms of evidence based-TA-type trading systems and you all would be surprised at how simple some of the stuff they use is. Simplicity is good not bad...as DeVinci said "simplicity is the ultimate sophistication".

one thing i've noticed about more simplistic methods is that they tend not to have very high winning percentages. I mean sure they can make money on average, but they are only right more often than not, rather than 90% of the time or more. i personally prefer high winning rates even if I have to give up some performance.

on the other hand if you're not over fitting the data (which is sometimes tough no to do), in my experience, more complexity will yield a higher win rate, but might have slightly lower returns overall.

yes the old school trend-followers tend to have lower hit ratios but their average winner dwarfs their average loser which is where the edge is derived...this is classic long vol trading. I tend to think that very high win rate strategies are dangerous because they are often short vol, mean reverting strategies that can produce big returns in sample but can blow up in epic fashion when we get the once a decade "crash of a lifetime"...or even just a bad blip on the radar. I similarly am always suspicous of high sharpe ratios. I am a discretionary trader, but I aim for a fairly low hit ratio and when i notice my hit ratio is getting too high I focus on making sure Im being aggresive enough with winning trades.

well sure, but on the other hand, you cant simply write off mean reversion just because of this, there were times were trend following did just as poorly, were people were constantly buying tops and selling bottoms, hoping for crashes that never came. they would have also blown up, by constantly cutting their losses. it also can take a long time for these events to occur, and sometimes by the time they occur you are already at a pretty big loss. Im personally not ok with with being down for several months in a row.

Ill add that i usually like to trade a high win rate, strategy AFTER it has a big drawdown, that seemed to be unprecedented in the data(of course i give in to my temptations every now and then). usually high win rate systems work well again after they shrug off weak levered hands. the chances of it having a big drawdown twice in a row are very slim. I almost never see them, especially in bigger markets.

Also I wouldn't write off high win rate strategies in general either. I would assume that just as there are high win rate strategies that use mean reversion in stocks( given the slightly negative aurtocorrelations), likewise there are probably high win rate strategies in commodities and fx, because they have slightly positive autocorrelations.

Also given the fact that these autocorrelations are the way they are, I must be quite suspicious of anyone shooting for mean reversion in commodities, and trend following in stocks.

jackbnimble Macro Arbitrage: jackbnimble:

Once again. Goldman Sachs traders are not building algorithms based on moving averages. neither is rentech, or Getco, or Jane Street, or ctc or Citadel.

Many PMs at those above HFs use momentum indicators extensively. I'm saying that not because I've spoken to an adequate sample of PMs at the above funds, much less one that would disclose their methods extensively, but because such indicators are rather popular among systematic macro traders. Kaveh Alamouti, Citadel's head of macro trading, once hired a bunch of trend following traders to manage money in the mid 2000s while at Moore Europe, including a famous 'turtle'. Rentech's Medallion fund also purely focused on TF strategies until the early 90s, at which point they transitioned into short term trading. Every single large macro fund I can think of, from old school Caxton to the more modern Brevan Howard, run an exclusive TF fund as well, and most of the time their entry signals involve using methods that are very similar to moving average crossovers, if not the same thing. Hell, even pure quant shops like TwoSigma and WorldQuant have managers running such strategies (publicly disclosed in the case of TwoSigma). I hope it is clear that you are thoroughly misinformed.

1) I have also spoken to people at some of these firms albeit not every single one you or I have mentioned. I assure you our conversations did not involve flag patterns or moving averages. In fact (though I am not saying they disclosed to me what their firm was doing) the vibe if you will, and the context and setting in which I was speaking to them indicated that they use the form of TA which I argue in favor of, the one where you test your ideas on empirical evidence to see if there is any basis in reality.

2)im pretty sure I acknowledged this when lbrest posted this. I understand that guys like druckenmiller have also spoken TA. But the vast majority of the profits (as in more than half) made at these firms are not made by blindly following trend lines. In fact with the exception of Goldman Sachs and Citadel, the prop shops I mentioned do not make the bulk of their profits off of blind TA. It seems to me that given the growing demand for tick data for example, and the fact that brokers sell their order flow to these firms, that they care deeply about testing their ideas, otherwise this information would be useless thus yielding no demand.

3)I never contested the history of rentech from the 90's, it seems to me they have chosen to pursue methods TODAY that are not related to momentum indicators. In fact Sebastian Mallaby mentions either rentech or shaw being able to make money exploiting anomalies. Anomalies by definition have to have empirical significance to be called anomalies. Moreover, It seems to me that rentech pursues a largely quantitative strategy given that he is a mathematician that hires other PHds in quantitative fields, keep in mind we had a major bull market during the 90's. DE Shaw has also explicitly mentioned the use of testing.

4) It appears I need to do more hw, with the the macro HF's. On the other hand the fact that most of the prop shops do not rely on TA for the bulk of their profits, is not wrong. But keep in mind I did not mention any macro HF's to begin with. Anyhow, I would probably wager that many of these hedge funds that use the momentum indicators, would likely do just as well, without the momentum indicator, as long as they are trading the direction of the market for whatever period that indicator is adjusted for. The reason I say this is because the markets that macro traders trade (fx, commodities and bonds), have slightly positive autocorrelations.

5) again i am not against the use of technical analysis, but I am against blindly selling a head and shoulders without first quantifying it and testing it.

6) the markets change, if moving average crossovers have worked for a long time in the 90's, they may or may not work now. Actually most of what you post in rebuttle to what my point was were examples from pre 2010. the only exceptions you mention were brevan and 2 sigma. Of course new opportunitities open up, and I have very good gut feeling that these two funds (especially 2 sigma) do not simply make decisions without testing them.

I'm afraid I did not bother reading prior posts and it seems we're kinda on the same page. I'm hardly an advocate of the 'eye-ball your way to financial freedom' version of TA either. For better or for worse, using basic robust methods coupled with novel statistical methods for validation is the new path that has been paved for systematic trading.

Macro Arbitrage jackbnimble: Macro Arbitrage: jackbnimble:

Once again. Goldman Sachs traders are not building algorithms based on moving averages. neither is rentech, or Getco, or Jane Street, or ctc or Citadel.

Many PMs at those above HFs use momentum indicators extensively. I'm saying that not because I've spoken to an adequate sample of PMs at the above funds, much less one that would disclose their methods extensively, but because such indicators are rather popular among systematic macro traders. Kaveh Alamouti, Citadel's head of macro trading, once hired a bunch of trend following traders to manage money in the mid 2000s while at Moore Europe, including a famous 'turtle'. Rentech's Medallion fund also purely focused on TF strategies until the early 90s, at which point they transitioned into short term trading. Every single large macro fund I can think of, from old school Caxton to the more modern Brevan Howard, run an exclusive TF fund as well, and most of the time their entry signals involve using methods that are very similar to moving average crossovers, if not the same thing. Hell, even pure quant shops like TwoSigma and WorldQuant have managers running such strategies (publicly disclosed in the case of TwoSigma). I hope it is clear that you are thoroughly misinformed.

1) I have also spoken to people at some of these firms albeit not every single one you or I have mentioned. I assure you our conversations did not involve flag patterns or moving averages. In fact (though I am not saying they disclosed to me what their firm was doing) the vibe if you will, and the context and setting in which I was speaking to them indicated that they use the form of TA which I argue in favor of, the one where you test your ideas on empirical evidence to see if there is any basis in reality.

2)im pretty sure I acknowledged this when lbrest posted this. I understand that guys like druckenmiller have also spoken TA. But the vast majority of the profits (as in more than half) made at these firms are not made by blindly following trend lines. In fact with the exception of Goldman Sachs and Citadel, the prop shops I mentioned do not make the bulk of their profits off of blind TA. It seems to me that given the growing demand for tick data for example, and the fact that brokers sell their order flow to these firms, that they care deeply about testing their ideas, otherwise this information would be useless thus yielding no demand.

3)I never contested the history of rentech from the 90's, it seems to me they have chosen to pursue methods TODAY that are not related to momentum indicators. In fact Sebastian Mallaby mentions either rentech or shaw being able to make money exploiting anomalies. Anomalies by definition have to have empirical significance to be called anomalies. Moreover, It seems to me that rentech pursues a largely quantitative strategy given that he is a mathematician that hires other PHds in quantitative fields, keep in mind we had a major bull market during the 90's. DE Shaw has also explicitly mentioned the use of testing.

4) It appears I need to do more hw, with the the macro HF's. On the other hand the fact that most of the prop shops do not rely on TA for the bulk of their profits, is not wrong. But keep in mind I did not mention any macro HF's to begin with. Anyhow, I would probably wager that many of these hedge funds that use the momentum indicators, would likely do just as well, without the momentum indicator, as long as they are trading the direction of the market for whatever period that indicator is adjusted for. The reason I say this is because the markets that macro traders trade (fx, commodities and bonds), have slightly positive autocorrelations.

5) again i am not against the use of technical analysis, but I am against blindly selling a head and shoulders without first quantifying it and testing it.

6) the markets change, if moving average crossovers have worked for a long time in the 90's, they may or may not work now. Actually most of what you post in rebuttle to what my point was were examples from pre 2010. the only exceptions you mention were brevan and 2 sigma. Of course new opportunitities open up, and I have very good gut feeling that these two funds (especially 2 sigma) do not simply make decisions without testing them.

I'm afraid I did not bother reading prior posts and it seems we're kinda on the same page. I'm hardly an advocate of the 'eye-ball your way to financial freedom' version of TA either. For better or for worse, using basic robust methods coupled with novel statistical methods for validation is the new path that has been paved for systematic trading.

lol glad we agree here.

Bondarb

2011 and 2012 were tuff time for the simple momentum guys (of which there are still many)...however, long-term many of these very simple strategies test out very well. I am skeptical of all these algo strategies now, even very complex ones that back-trest impeccably, because there is a ton of money in the space right now and there is too much data that I consider meaningless to current markets (ie testing data from 1992 to derive a strategy to trade interest rates in the new world of central bank interventionism, etc).

But in general I have worked with some people who are considered pretty cutting edge in terms of evidence based-TA-type trading systems and you all would be surprised at how simple some of the stuff they use is. Simplicity is good not bad...as DeVinci said "simplicity is the ultimate sophistication".

t think any strategy that was long vol lost money in 2011 and 2012 due to choppy markets. Kinda concerned about the significant AUM increase since 2008 as everyone wants to have a 'CTA put' in their allocation these days. No idea how that's gonna play out, but the behemoth funds are forced to focus on g-10 rates and fx, still leaving some room in other markets, particular EM which is an area I'm researching. At the same time, I believe most discretionary macro funds often maintain the same thematic trades as systematic funds, just that discretionary funds promptly enter and exit those trades earlier than systematic funds (i.e enter before the trend manifests and exit before the trend comes to an end but still ride most of the trend), so accordingly I believe that the CTA AUM increase could also impact discretionary macro, although to a lesser extent. Totally agree on simplicity and that's something I have to keep reminding myself cause its so easy to get carried away.

Macro Arbitrage Bondarb:

2011 and 2012 were tuff time for the simple momentum guys (of which there are still many)...however, long-term many of these very simple strategies test out very well. I am skeptical of all these algo strategies now, even very complex ones that back-trest impeccably, because there is a ton of money in the space right now and there is too much data that I consider meaningless to current markets (ie testing data from 1992 to derive a strategy to trade interest rates in the new world of central bank interventionism, etc).

But in general I have worked with some people who are considered pretty cutting edge in terms of evidence based-TA-type trading systems and you all would be surprised at how simple some of the stuff they use is. Simplicity is good not bad...as DeVinci said "simplicity is the ultimate sophistication".

t think any strategy that was long vol lost money in 2011 and 2012 due to choppy markets. Kinda concerned about the significant AUM increase since 2008 as everyone wants to have a 'CTA put' in their allocation these days. No idea how that's gonna play out, but the behemoth funds are forced to focus on g-10 rates and fx, still leaving some room in other markets, particular EM which is an area I'm researching. At the same time, I believe most discretionary macro funds often maintain the same thematic trades as systematic funds, just that discretionary funds promptly enter and exit those trades earlier than systematic funds (i.e enter before the trend manifests and exit before the trend comes to an end but still ride most of the trend), so accordingly I believe that the CTA AUM increase could also impact discretionary macro, although to a lesser extent. Totally agree on simplicity and that's something I have to keep reminding myself cause its so easy to get carried away.

as a discretionary macro trader i totally agree on the CTA/macro thing...discretionary macro can often just become faster-moving, quicker-trigger trend-following. You see that right now in trades like short AUD and long USDJPY which are both big macro trades and big CTA trades...last month when Nikkei snapped back violently both strategies on aggregate had draw-downs just that the CTA drawdowns were bigger. One difference right now is that i think alot of slower-moving CTAs got/are getting drilled in (and have exacerbated) the bond market sell-off whereas most macro was either flat or short.