Technical Analysis, from a Quant's Perspective

[MGJ]

Quote from quantzack:
Has anybody ever seen an Automated Technical Analysis trading system, which is non-subjective and rules-based, with clear entry/exit signals.

Depends on your definitions. There are lots of non-subjective, rules-based, unambiguous trading systems being traded every single day. The books written by Lars Kestner, or Thomas Stridsman, (or Tushar Chande, or Charles LeBeau, or ...) are brimming with such systems. There are over a thousand of these trading systems available for free on the wealth-lab.com website; click on "Published Chart Scripts" to see them and to try them out on the stocks of your choice, for free. Whether these meet your definition or not, is entirely up to you since you get to do the defining.

The first two Market Wizards books by Jack Schwager profile a good number of professionals who exclusively trade using non-subjective, rules-based, unambiguous signals from mechanical trading systems. You may enjoy reading what they have to say.

 

[quantzack]

Quote from MGJ:

There are lots of non-subjective, rules-based, unambiguous trading systems being traded every single day.

Would most traders define these non-subjective, mechanical systems as Technical Analysis?

 

[MGJ]

I'll wait a week or two, giving you a chance to explore the Wealth Lab website, look up those authors' books on Amazon, and perhaps do a bit of poking around on your own, either using Google or (gasp!) visiting a bookstore or library.

 

[landboy]

Quote from MGJ:

Mostly this is a disagreement about definitions. What's "Technical Analysis"? What is it that a "Quant" does? How much do they overlap? How do they differ?

For example, where do you put "Rule Based Mechanical Trading Systems?" Some would say "B". Others "A" and still others "C". Each of them are correct -- according to their definitions of quant and of TA.

Where do you put "Calendar Anomalies" like "Sell in May and Go Away" or "Buy 2 days before the end of the month and sell 2 days after the beginning of the next month"? A good case could be made for all four choices: A, B, C, D.

Where do you put Graham-and-Dodd "Value Investing"? At D? At A?

How about Financial Astrology?

When there is disagreement about "What is TA" and "What is Quant", there will necessarily be disagreement about which is more beneficial & profitable.

There really shouldn't be any overlap, it's pretty consistent if you read most hedge fund books, and not rely on the guy down the street saying he's a quant...

- Technical analysis is simply the study and belief that previous prices can determine future ones. That's it, you can get all complicated about it, but THAT's IT!!

- Quantitative Methods are regressive studies that take outside, variable inputs, that are correlated and to some degree predict prices. Ya, a Phd may say his random-walk theory may be quantitiative blah blah, but if he's using previous prices to predict future ones, it's simply TA...

 

[socalpt]

According to statistics, TA or QA only works
over 50% of the times; which means that you cut the loses and let the winners run. In the end, it doesn't matter the type of analysis you are using, one win will equals
at least 5 loses.

I know alot of traders made a living out of TA; some good examples are Bulkowski and Dan zanger.

Most auto trade setups today I believe are using TA methods of setup, and most FOREX traders also use TA almost exclusively on trades.

 

[whitster]

"According to statistics, TA or QA only works
over 50% of the times; which means that you cut the loses and let the winners run. In the end, it doesn't matter the type of analysis you are using, one win will equals
at least 5 loses. "

utter rubbish.

this is the trend following mantra (and that is ONE application of TA), but it is not the only form of TA.

simply put, it's rubbish. it is trying to apply one narrow methodology and assign it the status of universal truth

i have several TA based setups (i trade dow futures intraday - mostly scalps) that win over 80% of the time.

so spare me your fake statistics.

also, many TA setups can and do work (with positive expectancy) by setting the stop at a GREATER distance than the target.

one of the most persistent falsehoods is that the ONLY way to do it is have your target (if any) larger than your stop- the fabled "cut your losses short, let your winners run" philosophy./

that is applicable to SOME setups and SOME forms of TA, but not others.

it is no wonder that most traders fail, when they have such narrow, ideology (instead of fact) based thinking.

TA is the study of price, or its derivatives.

it clearly has value

the OP is correct, in that some forms of quantitative analysis ARE "TA with a pedigree", and nothing more

what matters it aht you have a methodology that has an edge, and that you use proper risk management to let the edge work out over time.

fundamental analysis works as well. obviously, not when intraday scalping an index though. i love using Fundies on my B&H and longer term swings. some of my best investments ever (600%+ returns on one stock) were almost pure fundamentals

TA is ultimately a derivative of OTHER's people fundies. obviously.

there are many ways to skin the market cat, and imo the first sign that somebody should be ran away from is when they try to apply a narrow ruleset and think it is the only way to be a successful trader.

there are MANY MANY ways to be successful

there are far more ways to be unsuccessful

 

[kiwi_trader]

Quote from socalpt:

According to statistics, TA or QA only works
over 50% of the times;

Not true or we would simply use random entries. And Van Tharp's acquaintance's experiment with random entries (which aren't so random if you check the methodology they actually used as they are determined partially by what happened last) doesn't give what a good technical trader would consider a good result.

 

[kiwi_trader]

This is an irrelevant issue.

If 5 different successful technical traders (not analysts) use 5 different methods because most of us find methods that work for us why would we expect a person who doesn't believe that any of them work to be able to prove that they do?

All that matters is "can you construct a strategy, plan, risk allocator, that is compatible with your beliefs and you can consistently make money with?"

 

[mokwit]

Academics should find other things to look at.

The idea that we are out there looking for the text book patterns they test and entering with no thought to context and overall market conditions is the basic flaw in their argument. Reading price action is a craft based on knowledge of how socks are moved up and you are not going to duplicate years of learning with an algorithm for a double bottom. Secondly the division between technical and fundamental is meaningless and demonstrates overall ignorance, many use multiple inputs of which TA and FA are just a couple.

I just get so fed up with the drivel academics come out with WRT the market. Thee only thing they are demonstrating is why they are low paid and why the chances of making it to partner at Goldman with their research are so low. The one that made it did not even know that bids get pulled in a crisis.

For the benefit of academics reading this, Simons and Shaw are just outliers according to your empirical methodology.

 

[Mike805]

Quote from kiwi_trader:

All that matters is "can you construct a strategy, plan, risk allocator, that is compatible with your beliefs and you can consistently make money with?"

Thanks for this concise statement.

To the OP and others,

Take a look at Edwards & Magee works on TA. It's useful because it provides a framework for making objective decisions. It can be as simple as "price held $10 at time X, hence there is a good chance it will hold $10 now". The "edge" exists in the fact that you can sell at previously defined psychological price points, say for example $9 or $15.

From what I've witnessed, a discussion of TA usually will evolve into the prediction paradigm, which is bullshit IMO. If you read Magee, you'll notice that past data allows for current price interpretation, NOT accurate prediction. To some, this is a method of forecasting, to me, this is a matter of defining "edge" in a trade.

I ask anyone here: why the f--k do you care about predicting the future? Do you honestly think it will make you money? I'll argue it won't.

For me, the "edge" some here speak of is inherent in the trade. Your trade(ing) has "edge" when your positioning (i.e. entry) into the market allows for greater profit rather than loss. One can use any method they want to achieve this goal. TA allows it. MP allows it. As always, execution is what matters 90% of the time.

Mike