To what instruments is tech. analysis applicable?

iliavko

Newbie
Messages
6
Likes
0
Hey everyone! I have a nooby question for you,

To what financial instruments can I apply technical analysis? I am still new to this and I've been reading TA for Dummies as my first TA book and it seems to be focused on stock, I haven't seen any other instruments being used for examples.

What I mean by the question is, it sems highly applicable to stock, but for example, not so applicable to commodities futures since these move due to fundamental reasons or I am wrong? Is there such a thing as an "over\under-valued futures contract"?
Can you apply TA for options? Or Forex?.. Spreadbetting? Anyways what instruments you think it works best with?

-Can you apply TA techniques that you use, say, for stock, on futures markets? or index trading?

-What about timeframes? The TA techniques I've seen so far seem to be applicable on a few weeks timeframes, but what about those who trade minut-by-minute? Or maybe there are specific tools applicable to these short timeframes?

Ayways, hope I made myself clear here, I'd appreciate your help!

Thanks guys!
 
If by "technical analysis" you mean the analysis of price movement, it is applicable to any instrument which moves according to the law of supply and demand. If by "technical analysis" you mean indicators, patterns, etc, that's another matter.

If you're interested in how the analysis of price movement is exemplified in futures trading, click the first link in my signature.

Db
 
Thank you for the quick reply!

Yes, by TA I mean indicators, patterns, etc. So by learning how, say, Elliott wave works with stock, I won't be able to apply those principles to, say, FX. Am I correct? If yes, then it kinda makes TA learning a bit messy or maybe the right word is "slower" as there is no useless learning, but it seems that you must have a clear idea of what exactly you want to trade and how (I mean your style) and only then you learn TA so it fits your criteria. I am right?

I'll check your link in a sec! Just onto something right now.

Thanks a lot for help!
 
I don't pay any attention to indicators or geometric patterns, so I can't help you there. As for the rest, the attached may be helpful:
 

Attachments

  • Beginner's Guide to Trading Price.pdf
    811.3 KB · Views: 519
Just to add to dbphoenix's comments - in terms of timeframes. Yes, technical analysis can be used on any timeframe. However, generally speaking, the higher timeframe the more reliable the patterns. By reliable I mean that the technical pattern's has a higher success for predicting the future direction.

Thus a head & shoulder pattern on a weekly timeframe is likely to be more reliable than the same pattern on a 5-minute timeframe.
 
Generally speaking, TA can be applied to any market where you have decent volume - stocks, rates, commodities, forex. I'd probably avoid using it for something like options where volume can be patchy, though applying it to the underlying market is just fine.

I'd also agree that as you move up the time scale away from the very short term, the patterns tend to be more stable. In the very short term there tends to be a lot more "noise".
 
Based on statistical analysis, widely traded financial instruments are essentially "random walks", which means they are unpredictable. Therefore, technical analysis that just looks at the price history of a single instrument doesn't work very well. Daily price movements are simply too random.

I trade a cointegrated PORTFOLIO of stocks/bonds/commodity ETFs which I carefully assemble so as to minimize the random component and maximize the deterministic (predictable) component. This is known as statistical arbitrage. By applying what is essentially a Bollinger band (mean reversion) statistic to a cointegrated portfolio, I've developed a profitable trading system.
 
Based on statistical analysis, widely traded financial instruments are essentially "random walks", which means they are unpredictable.

Actually, it doesn't mean that at all. Again.

That something is unpredictable does not mean that it's random.

As for your continued boasts of your allegedly-profitable "cointegrated portfolio", none of that means much unless and until you explain how you do it.

Db
 
As for your continued boasts of your allegedly-profitable "cointegrated portfolio", none of that means much unless and until you explain how you do it.

Why are you so hostile to new members? At the rate we're understandably losing people, at the moment, we should surely welcome them. I, for one, am potentially interested in hearing more from IFeelFree, if and when he's ready to share it and hasn't been alienated by your rude and dismissive comment.
 
Actually, it doesn't mean that at all. Again.

That something is unpredictable does not mean that it's random.

As for your continued boasts of your allegedly-profitable "cointegrated portfolio", none of that means much unless and until you explain how you do it.

Db

You are quite right in saying that if something is unpredictable, that doesn't mean it's random. However, that's not what I said. I'm saying that if something is random, it is, by definition, unpredictable. Widely traded stocks, for example, are very close to being a random walk, which means they are unpredictable using the tools of statistics. (If that were not true, everybody could get rich using statistics to trade those stocks.) Cointegration is used to find portfolios of stocks, ETFs, etc., in which the random component is minimized, and the deterministic (predictable) component is maximized.

As for explaining my approach to trading a cointegrated portfolio, I'd be happy to explain further. I've written software which tests groups of price histories for cointegration, using the Johansen test. I've found a cointegrated triplet of leveraged ETFs which I'm currently using for trading. After using the Johansen test to obtain the correct portfolio weightings, I use a Kalman filter to dynamically update the 3-component mean spread, and the deviation from the mean spread which gives the buy/sell signals, similar to a Bollinger band (except without the lags inherent in moving averages and moving standard deviations). I use backtesting to obtain optimal threshold levels for entering and exiting a trade. If any of this is not clear, I'd be happy to explain further.

Currently, I'm trading doing about 1 trade per week (weekly data) and getting >50% average yearly return, with a maximum drawdown of 9%. I'm experimenting with adding a cointegrated VAR model, and that appears to significantly improve performance.
 
You are quite right in saying that if something is unpredictable, that doesn't mean it's random. However, that's not what I said. I'm saying that if something is random, it is, by definition, unpredictable. Widely traded stocks, for example, are very close to being a random walk, which means they are unpredictable using the tools of statistics. (If that were not true, everybody could get rich using statistics to trade those stocks.) Cointegration is used to find portfolios of stocks, ETFs, etc., in which the random component is minimized, and the deterministic (predictable) component is maximized.

You're being disingenuous. Price movement is not random. Therefore, whatever conclusions you may draw about their unpredictability are based on a false premise. Whether or not one can get rich by using statistics is another issue. People believe that they can get rich by using the right settings with the right indicators, but that doesn't make it true.
 
You're being disingenuous. Price movement is not random. Therefore, whatever conclusions you may draw about their unpredictability are based on a false premise. Whether or not one can get rich by using statistics is another issue. People believe that they can get rich by using the right settings with the right indicators, but that doesn't make it true.

I'm not being disingenuous. I sincerely believe what I'm saying. (You could argue that I'm wrong, in which case you would need to present evidence of that.)

As for "Price movement is not random", see the "Random Walk Hypothesis", for example:

https://en.wikipedia.org/wiki/Random_walk_hypothesis

When one looks at the distribution of, say, daily price changes of, for example, a widely traded stock, or a stock market index, it is close to being a normal (Gaussian) distribution, which is the distribution you would expect of a random walk. The distribution of price changes is not EXACTLY a normal distribution (it has "fat tails"), which is why I said it was "very close to being a random walk". Thus, there appears to be a small deterministic trend in stock prices. That's why long-term buy-and-hold investors in the broad stock market can expect to have a positive return over time periods of 20 years or more. For the short-term trader, however, returns are essentially random, having a normal distribution with fat tails.

Note that what I'm saying applies only to statistical analysis of the price history of a single trading instrument in isolation. Many traders successfully use additional information, outside of statistical analysis, to trade profitably. Also, statistical arbitrage combines multiple trading instruments to trade profitably.
 
People believe all sorts of things. That doesn't make all or any of them true. And I'm well aware of the "random walk hypothesis". And price movement is not random. And, no, the daily price changes of an instrument are not a normal distribution. If you want to believe they are, great. But please don't advance the position as established fact.
 
People believe all sorts of things. That doesn't make all or any of them true. And I'm well aware of the "random walk hypothesis". And price movement is not random. And, no, the daily price changes of an instrument are not a normal distribution. If you want to believe they are, great. But please don't advance the position as established fact.

See:

https://en.wikipedia.org/wiki/Share_price

"In economics and financial theory, analysts use random walk techniques to model behavior of asset prices, in particular share prices on stock markets, currency exchange rates and commodity prices. This practice has its basis in the presumption that investors act rationally and without biases, and that at any moment they estimate the value of an asset based on future expectations. Under these conditions, all existing information affects the price, which changes only when new information comes out. By definition, new information appears randomly and influences the asset price randomly."

However,

"Empirical studies have demonstrated that prices do not completely follow random walks.[1] Low serial correlations (around 0.05) exist in the short term, and slightly stronger correlations over the longer term. Their sign and the strength depend on a variety of factors."
 
Returning to a more practical level.

TA is a useful indicator towards what the market thinks about and is doing about a given instrument. It doesn't comment directly on the currencies or national economies involved in a given forex exchange rate, it doesn't tell you anything directly about a company's operations or management or even their balance sheet. Its more an opinion poll than a scientific valuation.

The simplest question in TA is, is price rising or falling? Prices that have been rising steadily and for an extended period are in an uptrend: trends are more likely to continue rather than reverse or stop. Trading with the trend makes this game easy and profitable.

But don't ignore the time issue which can invalidate trend trading. If a forex price has been in an uptrend for 68 weeks and rose on 7 of the last 10 days, that doesn't mean it will definitely rise in the next 15 minutes: but it probably means it will rise in the next 10 days.

TA patterns like head-and-shoulders and continuation patterns and swing lows and shooting stars are embellishments which might or might not help in your trading. They probably won't help if you lose sight of trend.
 
"This practice has its basis in the presumption that investors act rationally and without biases, and that at any moment they estimate the value of an asset based on future expectations."

If you ever see anything that presumes rationality among investors be very skeptical. Even in academia there is a ton of research which documents "anomalous" price behavior - stuff that isn't rational.

To the larger point of the thread, Technical Analysis can be said to attempt to recognize and utilize the not-fully-rational and efficient nature of the markets.
 
See:

https://en.wikipedia.org/wiki/Share_price

"In economics and financial theory, analysts use random walk techniques to model behavior of asset prices, in particular share prices on stock markets, currency exchange rates and commodity prices. This practice has its basis in the presumption that investors act rationally and without biases, and that at any moment they estimate the value of an asset based on future expectations. Under these conditions, all existing information affects the price, which changes only when new information comes out. By definition, new information appears randomly and influences the asset price randomly."

However,

"Empirical studies have demonstrated that prices do not completely follow random walks.[1] Low serial correlations (around 0.05) exist in the short term, and slightly stronger correlations over the longer term. Their sign and the strength depend on a variety of factors."

Beginning traders commonly make the same mistake: they study books and similar printed material rather than the market. Those who take the former route often/usually adopt whatever set of beliefs are propounded by the book(s). Those who choose the latter route develop strategies that are tied to market realities rather than theoretical and philosophical constructs.

Rather than "see" some quasi-academic work, observe the market. Apply the scientific method to whatever hypotheses occur to one. Develop a consistently-profitable trading plan. Then adhere to it.

It's that simple.

Db
 
Beginning traders commonly make the same mistake: they study books and similar printed material rather than the market. Those who take the former route often/usually adopt whatever set of beliefs are propounded by the book(s). Those who choose the latter route develop strategies that are tied to market realities rather than theoretical and philosophical constructs.

Rather than "see" some quasi-academic work, observe the market. Apply the scientific method to whatever hypotheses occur to one. Develop a consistently-profitable trading plan. Then adhere to it.

It's that simple.

Db

I've been investing and trading for 30 years, so I don't exactly consider myself a "beginning trader". As I have a strong math and programming background, my preference is toward algorithmic methods. Obviously, there are many other successful methods of trading. To each his own.

My effort has been to use sound statistical approaches to trading, so as to avoid being "fooled by randomness" -- seeing elephants in the clouds, that is, perceiving patterns that are really the product of randomness. We all have have a tendency toward cognitive biases, so we have to adopt a disciplined approach to trading. As you say, "Develop a consistently-profitable trading plan. Then adhere to it."
 
And I've been trading and investing for 40 years, but that's not pertinent. The fact remains that we develop cognitive biases early. We may even approach the market with our biases already intact, e.g., trading is a "war". Those biases, however, are in us, not in the market, and if one intends to puzzle out the market, he's better off studying the market than studying monographs.

One can develop a consistently-profitable trading plan believing that markets are random. One can also develop such a plan believing the opposite. One can even develop such a plan that is based on planetary alignments. But a belief is not a fact, and to present a belief such as that markets are random as a fact does not reflect the intellectual rigor that mathematicians like to claim for themselves.

Db
 
Top