Don't Be Fooled By Randomness

[DbPhoenix 42]

That's what I thought.

 

We eagerly await your explanation of how you managed to repeal the law of supply and demand. If you need more than five days, just let us know.

[TheDude 10]

Again, wrong answer. Please try harder.

 

You're thinking is off. It seems stuck in fact.

 

Here's a clue:

 

One way is based on perceived supply and demand,

 

The other isnt.

 

Another clue:

 

May be there are other ways to make money than the only one you can think of? Surprising that after all these years of analysis of the markets, you are only aware of one way of making money. Wasted years perhaps?

 

 

I will hold my tongue/fingers until the weekend....

 

Please try and think a little harder. I have every confidence in you

 

[Gekko78 10]

TA, however, isn't about patterns and indicators or even charts. TA is about price movement. All the rest of that stuff came later.

 

I guess it depends on what you are referring to here.......if buy TA you are referring to Tape reading then yes I agree that was , and still is , the best TA out there.....

 

Tape reading is the ONLY thing IMO that can most accurately predict what is about to happen.

 

MACD , Stochastics , RSI and all that other stuff are not TA , that is math calculating the past.......thats it.

[DbPhoenix 42]

I guess it depends on what you are referring to here.......if buy TA you are referring to Tape reading then yes I agree that was , and still is , the best TA out there.....

 

Yes, though before the "tape", there were reports and logs, such as were used by Homma. The progress of price can be tracked in a number of ways other than by charts and the tape.

[SunTrader 10]

I doubt seriously that any tape reading old codger would take kindly to being called a Technical Analyst.:puke:

[DbPhoenix 42]

I doubt seriously that any tape reading old codger would take kindly to being called a Technical Analyst.

 

I've been called worse.

[Peterthemonkey 10]

Without definitions arguments are useless.

 

To me, and I believe most, TA is the study of trading instruments price behavior through the use of price and/or volume and/or time, with many sub-disciplines: tape reading, chart analysis, cycles ..................... FA analysis is done through the use of market economic conditions as they apply to markets, companies, persons and such, again with it's many sub-disciplines characterized by individual field of specialization.

 

If we accept the general definition used by most, fine ........... if not, fine. Argue a specific point for validity, without wandering all over the place.

 

To my understanding, random is anything that exists without a cause, there is no causal factor that made it be. Thus, any effect that is a product of a cause is not random, but it could easily be beyond the ability of anyone to predict it with 100 % certainty, as is the case with price behavior.

[FFTrader 10]

Without definitions arguments are useless.

 

To me, and I believe most, TA is the study of trading instruments price behavior through the use of price and/or volume and/or time, with many sub-disciplines: tape reading, chart analysis, cycles ..................... FA analysis is done through the use of market economic conditions as they apply to markets, companies, persons and such, again with it's many sub-disciplines characterized by individual field of specialization.

 

If we accept the general definition used by most, fine ........... if not, fine. Argue a specific point for validity, without wandering all over the place.

 

To my understanding, random is anything that exists without a cause, there is no causal factor that made it be. Thus, any effect that is a product of a cause is not random, but it could easily be beyond the ability of anyone to predict it with 100 % certainty, as is the case with price behavior.

 

Peterthemonkey put out a good point - you need to define it.

 

If anyone here have studied information theory - see: http://www.framingham.edu/~dkeil/dscs-chaos.pdf

 

Go to slide 18 - note that the famous ratio of Pi ... is by definition not random. Why? Because it can be generated by a short algorithm (of mathematics).

 

In other words, Burton Malkiel famous coin tossing experiment and mathematically adding or subtracting a constant is also using an algorithm (of addition/subtraction) and thus by definition not random mathematically?

 

Malkiel may have underestimated that in Information Theory (Google Claude E. Shannon if anyone cares), there are a mountain of knowledge and foundation concerning coding, telecommunication bandwidth, channel capacity, etc. that builds on the idea of randomness ...

 

Yeah, just leave it to the economist and their real understanding on mathematical principles? :doh: :missy: :doh:

[TheDude 10]

So statistical and quantative analysis is TA is it?

 

e.g. Im a stat/quant fund. (and you think thats TA lol)

 

my number crunching suggests i should go long xyz, short abc. I do so. when i close the trade, based on more number crunching, xyz is the same price, so I scratch. abc is also the same price, so that too is scratched.

 

YET i still make money on the trade.

 

How come?

 

Come on Phoenix. You know everything about everything. Tell me how I made money using a branch of TA where there was no price movement?

 

 

This I got to see!....

 

Ringside seats folks......

 

How to polish a turd coming up.......

 

Times up Phoenix. Looks like you FAILED.

 

Surprising since TA is THE ONLY way to analyse the market, and TA is all encompassing unless it's FA (according to you).

 

Anyway, 3 ways I could have made money despite scratching the long and short trade:

 

1. The obvious one. Dividends. The short position would have been a hedge of course.

 

2. Also quite obvious - options, where the stock is the hedge, ie short calls on the long, short puts on the short stock. I'd pick up time premium. Time decay is neither FA, TA, or any A. It's a F. F is for FACT that options have this characteristic.

 

3. Stock lending. ok a bit of a trick one this, but still valid. I could have loaned out my long position to another hedgefund who wished to short the stock. I would have charged them for that of course. The short was a diversion. But as the expert guru on TA, Im sure you are used to diversions, and things going up when the chart says down etc.

 

There are more....

 

I hope you see now that not everything is TA. Hopefully now you see that the world is not black or white (TA/FA) but there are many, many different themes, colors around us.

 

Perhaps a little humility and objectivity wouldnt go a miss in future?

 

Lets see how you weasel your way of this.....

[jwat 10]

Some studies have shown that a high percentage of chart patterns may be simply random lines on a piece of paper. Burton Malkiel in A Random Walk Down Wall Street cited an experiment his students participated in, constructing a hypothetical stock chart. Each day they flipped a coin, plotting heads as a 1/2-point gain and tails as a 1/2-point decline. The resulting chart from these random coin flips displayed all the classical patterns such as head and shoulder formations, flags, pennants, triangles, etc. There were even indications of cycles.

 

There is a great site where you can test out technical analysis by comparing it to randomness -nonrandomwalk. You have to love the name. At the site, you practice trading blindly (you don't know what stock you are trading or what year you are trading it) on historical data and then compare how you did to a set of traders randomly trading the same stock. The results are often humbling.

 

With regard to Malkiel, his experiment has a fundamental problem. Just because coin-flipping can produce classical patterns, that doesn't mean that the patterns in the market are random. One would have to look at the frequency with which patterns appear in coin flipping versus actual data. To my knowledge he did not do that. People who have (Andrew Lo) have found that chart patterns occur with demonstrably more frequency than would be expected if markets were just coin-flipping.

 

Consider another example from a standard probability class. Imagine a thousand apes randomly typing on a thousand type-writers for an infinite amount of time. What is the probability that they will eventually, in combination, type the complete works of Shakespere? The probability is one. It is an interesting mathematical proof. Does this mean that the works of Shakespere were random?

[Seeker 11]

Obviously there is a lot of semantics going on.

 

Technical analysis is the analysis of the technicals. What are the technicals? They seem to be price, volume etc. Taken as that, chart analysis is TA, so too ismarket profile.

 

Generally those who want to divide and exclude certain parts, say market profile, are doing so to further their own agenda, to make themselves stand out. For example, if Steidlmayer wants to promote his method, he may wish to state that his method is different from all other methods, or most common methods, i.e. TA. Is it a form of analysis on the technicals? It is to me. If the technicals involve price, then how could one argue it's not. But it's not worth getting upset about just because you take the word to mean something. There's a good reason why Steidlmayer would want to say it's not TA, the Dude. Simply believing him without understanding his motive doesn't make for a good argument.

 

I'd also disagree with DBPhoenix about the word random. In some sense isn't almost everything deterministic? The coin toss for example is deterministic, it depends on laws and forces. However, we model it as random because we cannot know pecisely enough the information that we would need to know to decide the outcome with certainty. Therefore this applies to trading too. It's irrelevant whether there are causes, of course there are. But scientifically, mathematically it is random still because you cannot with certainty tell me exactly what will happen.

 

Therefore DBPhoenix, although I agree with you on TA, I'd have to say that you're wrong on the statement

"The outcome of any given trade is unknowable. That doesn't make it random. "

 

It exactly DOES make it random.

Edited March 15, 2013 by Seeker

[gianno 10]

Another post in the blog @equtrader mentioned where it is explained how TA and backtesting lead to randomness but it falls short of any suggestions for solution.

[natureboy1618 10]

TA needs to be objective. Period. Otherwise its unverifiable.

 

Here's the mindbender for me: It is said price is somewhat random {Andrew Lo, A Non-Random Walk Down Wall Street}. It is said many technicians can't tell a real price chart from a randomly generated one. The bender: check out the Wikipedia page for Ito Calculus.

 

There is a graph on an Ito Integral of Brownian (random) motion. Tell me that's not a Bullish Butterfly pattern in the middle (the low) of the chart at T 1.5. I find these Butterflies (also called M7s) in other time series that could be easily thought of (if inaccurately) as random ... such as total solar irradiance, crime rates, barometric pressure, % of men with beards, etc..

 

The weird thing is that M7 patterns do NOT have a random outcome.. they precede reversals in the time series 60%+ of the time (from what i have measured, far from the population).. even in 'random' data!

 

Someone please explain that to me

[DbPhoenix 42]

Depends on how one defines TA. If it is defined by the use of indicators (post-1950), one will achieve a certain set of results. If it is defined via patterns (post-1930), one will likely achieve another. If it is defined as it was originally as the analysis of the imbalances between buying pressure and selling pressure (post-1750), then it should come as no surprise that the "butterfly pattern" precedes reversals since the "pattern" is essentially a double bottom or double top. These in and of themselves will often precede reversals. If they also occur at support or resistance, the instance will be even higher.

[TheDude 10]

It's all utter bull excrement. It does not matter one jot.

 

One does not make money from technical ANALYSIS, you make money TRADING. Those are 2 verbs with very different meanings.

 

I can guarantee you this:

 

If you spend your years becoming an expert in analysis, you may be a very knowledgeable analyst, but still totally unable to make any money because you spent all your time studying market history, not understanding how to act NOW. ie trading.

 

90% fail because they dedicate 80-90% of their time to TA and not trading skills.

 

Butterflies, tripple tops, box, magic lines, pennants, flags, blah blah blah. What a joke!

 

Please folk - looking at shapes on a chart is basically like reading tea leaves. It also carries as much intellectual weight in true trading circles. If I notice patterns of tea leaves or moon cycles and correlate them to market moves, is it coincidence (no matter how loosely defined they are), or is it proven fact I can call a statistical edge?

 

PLLLEEEEEEAAAAASSSSSSSEEEE!!!!

 

Get a grip folks.

[natureboy1618 10]

....the whole point is - a butterfly is testable. No its not TRADING. Its a price behavior. This thread is about randomness in price.

 

Trading is a whole other activity than analysis - totally. You can make money trading with zero statistically verified analysis. Its uncommon, guys have done it.

 

At any rate, the whole point of studying TA is to find the edge to act on. As far as trading circles go, i've spoken with several people who trade successfully for a living and they do take chart patterns seriously. Its not tea leaves man. Andrew Lo - MIT - .... studied Double Top/Bottom and others... Best grip you can have is a repeating price behavior..

 

Bull excrement is a harsh vote.. In the light of modern research, i'd say its an uninformed vote. TA can work - TA alone won't make you money, yeah that's right.

[natureboy1618 10]

... a double bottom or double top. These in and of themselves will often precede reversals. If they also occur at support or resistance, the instance will be even higher.

 

i know, but its weird to find the same patterns in 'random data' ... this has zero to do with making money. Check. Its just thinking material. If you could show that patterns with say 72% probability leading indication are in random data, what does THAT say about randomness... for instance? Nothing is ever totally random

[DbPhoenix 42]

i know, but its weird to find the same patterns in 'random data' ... this has zero to do with making money. Check. Its just thinking material. If you could show that patterns with say 72% probability leading indication are in random data, what does THAT say about randomness... for instance? Nothing is ever totally random

 

I don't find it strange to find patterns in random data. You don't think those people who see Jesus in their morning toast are onto something, do you?

 

But take care not to confuse "random" with "unknowable". Nor can one ignore probability. The outcome of a particular trade may be unknowable, but that doesn't mean that it's random if one's testing has determined that the probable outcome is X. That is, after all, the basis of the edge.

[natureboy1618 10]

You don't think those people who see Jesus in their morning toast are onto something, do you?

 

 

 

I did until you said something...

 

it seems like a random model is applied a disproportionate amount of the time as a crutch.. an approximation.. in quant finance. My posts in this thread are completely on an academic level - curious how folks around here think. Its not so much that patterns are present in random data, more that certain patterns in random data could hold an element of nonrandomness that strikes me... its not surprising to me anymore, either... a thread on randomness, had to offer my

[steve46 12]

This isn't particularly difficult to understand...markets cycle from random to non-random continuously, AND patterns both random and non-random replicate all the time, what matters is the significance of the pattern in terms of time (timing) not that it exists.....

 

"Mathematics of Technical Analysis"....Cliff Sherry

 

Its been a while but this was one of the texts that seemed to get it right

[SIUYA 11]

i am sure there are better places than this thread for it, but strangely I find it apt

 

PIMCO | Investment Outlook - ?A Man in the Mirror

[BlueHorseshoe 10]

 

Someone please explain that to me

 

Hello,

 

It's fairly simple really: perfectly random behaviour is highly predictable. The probability of a long sequence of similar events is very small. You can bet against the continuation of such sequences as they unfold.

 

Please do note that I am saying the explanation to your question about randomness is simple, and NOT that trading is simple!

 

Regards,

 

BlueHorseshoe

[TheDude 10]

No diagnosis,no prognosis,no prognosis,no profit.

 

Indeed.

 

However you must make sure youre diagnosing the correct information, not something that looks visually easy to grasp, easy to understand, but in fact offers zero edge.

 

Trading is about numbers. It's not about visual patterns. Thats what art class is for.

[TheDude 10]

....the whole point is - a butterfly is testable. No its not TRADING. Its a price behavior. This thread is about randomness in price.

 

1. Trading is a whole other activity than analysis - totally. You can make money trading with zero statistically verified analysis. Its uncommon, guys have done it.

 

2. At any rate, the whole point of studying TA is to find the edge to act on. As far as trading circles go, i've spoken with several people who trade successfully for a living and they do take chart patterns seriously. Its not tea leaves man. Andrew Lo - MIT - .... studied Double Top/Bottom and others... Best grip you can have is a repeating price behavior..

 

Bull excrement is a harsh vote.. In the light of modern research, i'd say its an uninformed vote. TA can work - TA alone won't make you money, yeah that's right.

 

1. Thats called luck

 

2. TA offers no edge. How can you get an edge from something that cant be defined in exact enough ways to measure?

 

This one keeps coming up, yet none of the TA fans are ever able to provide the evidence of statistical proof that TA offers an edge. There is not one study that has proven this.

 

If TA did offer an edge, trading would be easy due to the ease of learning TA.

 

If all you needed was to spot double tops/bottoms, flags etc, with the 'correct' associated volume pattern, all you folks would be rich.

 

However, given the amount of posts and time some here spend, my money is that they are not rich, otherwise they'd be out spending cash on fun stuff!

 

Anyways...back to the topic.... randomness is simply a word we use for something we dont understand. I took that from Taleb.

[steve46 12]

This thread (particularly the last few posts) reminds me of a recent commercial...the scenario is that an adult sits at a small (very small) table, with his knees up around his neck to emphasis the difference between him and the children sitting all round him...and he asks them a question...("what would you do with a million dollars?") and as one might expect they answer with typical unrealistic comments like...."I'd buy an island made out of candy" and "yeah and the sand would be made of sugar"....etc

 

With apologies to Taleb and those present who like to see their comments in print, but haven't seen fit to actually look up the definition...here is the citation;

 

RANDOM.ORG - Introduction to Randomness and Random Numbers

 

I am pretty sure Taleb knows this, not so sure about the rest of you...

 

and for the record (again) markets cycle from random to non-random, and for me at least what matters isn't whether price patterns are one or the other but the timing. Why because if you can determine the difference between random and non-random AND you know where the critical times are during the trading day, its possible to obtain an edge from that knowledge...(referring back to Sherry's book).