Random Walks & Onemind

[ducati916]

The scientific method requires only one example to disprove a theory.
The theory is assumed to be valid if no proofs to the contrary can be demonstrated.


Wow,my mathematician father warned me that economists were worse than physicists with their profligate use of the term proof.

I see why now.
Whilst I agree with the first statement, the second worries me.

So I guess that god exists as I cannot fond a counter example to his existence.

Well that is settled then. Happy easter!


Do you guys really believe something to be true merely because there are no proofs to the contrary? This guy was joking right?

I know this is an old thread, but that is really scary stuff people....

pedro

The theory promulgated is of course from Karl Popper, and is commonly referred to as falsification

I see that you agree with the first statement. The second statement is really nothing more than an inversion of the first statement. If, I can find no examples [evidence] to disprove the hypothesis, then, logically, by definition, it is true.

Your example displays clearly your muddled thinking on the topic. Let's restate the example correctly, sticking with your hypothesis.

So I guess that god exists as I cannot fond a counter example to his existence.

Simply invert your hypothesis:

H1: God does not exist.

Now, simply find one example of evidence, that is not based on hearsay, faith, that demonstrates God's existence, and God is proven to exist.

Of course the argument that rages in theological circles and beyond, is, does faith, constitute evidence?

jog on
duc

 

[ducati916]

Did I miss something?
Maybe the Random Walk theory is correct - maybe it is not.
But the notions of proof to which Ducati916 subscribes are just silly.


I have a question.

Does the proximity in time to the point at which a trend or pattern is observed have a correlation to its predictive power?

In other words, are, for example, double tops and wedges considered more reliable indicators in day trading than they are when used for trading over weekly or monthly timescales?

Should I start a new thread to ask this? I have a bunch more questions to bug you with.

By the way, anybody who calls himself Ducati916 has got to be a basically O.K. kind of a bloke if a little too much of a risk taker.

That a trend exists, falsifies the theory of a random walk. A random walk is based mathematically upon independence of the data. Stock prices are not independent, ergo, stock prices are not a random walk.

Nothing more really need be added. Timeframes are simply a red herring to the central question; viz. random walks.

jog on
duc

 

[ducati916]

Re: trends adn shapes -scaling properties?

Disregarding randomness, and imagining that I subscribe to technical analysis as a predictive tool, are trends in the more recent shorter term more useful than trends discerned over longer periods?

What I mean is - although line charts exhibit scaling properties (monthly, weekly, daily and intraday charts are indistinguishable without their time and price marked on them) can the same be said of these entities we call trends and shapes.

So is a double top considered as valid if the peaks of the double top are months apart rather than minutes apart?

Is there a difference to the chartist?

Agreed charts are fractal in appearance.

Yes, there are very real potential differences of the stock concerned. Take a simple example. The common stock traded at $40/share on April 2008, with 1,000,000 shares outstanding, traded down to $10/share, and while trading at that price, purchased back 500,000 shares, so that now there are 500,000 shares outstanding.

Thus, at $40 share, assuming the stock trades back up, we have half the capitalization that we had previously...a significant difference, and depending on earnings, potentially a far cheaper value on the stock.

Is there a difference to the chartist...to a good one, who is cognizent of the difference, yes. To the poorly informed, none whatsoever.

jog on
duc

 

[pedrochemical]

Is there a difference to the chartist...to a good one, who is cognizent of the difference, yes. To the poorly informed, none whatsoever.

This is more what I was getting at.

Now don't laugh, but I have a lot of data and am in the process of taking a Poincare Section.
My computer is not fast enough, however, to draw the pretty pictures.
The progression through phase space (vector field style) of what I am seeing is what I was expecting to see. This makes me nervous as an experimenter, but even though the numbers are as yet muddy, I hope to see the difference between randomly generated sequences and historic sequences.

I shall let y'all know what happens.
Pretty old school maths but applied like I am trying it - could be interesting if not novel.

Thanks for the input.

 

[pedrochemical]

As for Popper - I could imagine him saying that if a counter example does not exist, the theorem becomes a proof.
But we seem to be saying here is that if we "can't find one yet" the theorem is a proof - for me it remains a conjecture/hypothesis.

Have I missed something?

 

[pedrochemical]

That a trend exists, falsifies the theory of a random walk. A random walk is based mathematically upon independence of the data. Stock prices are not independent, ergo, stock prices are not a random walk.

Nothing more really need be added. Timeframes are simply a red herring to the central question; viz. random walks.

jog on
duc

Not so sure - according to this -

https://www.aussiestockforums.com/forums/showthread.php?t=9680

A randomly generated chart has trends too - "tradeable" trends at that, if you believe in Chartism.

As it happens I think I agree with you in principle though.

Also, time-frame in regards to psychology could be important in my view. A human is known to flinch in the short term and act more perspicaciously over longer time-frames.

When we see a wasp out of the corner of our eye we involuntarily move rapidly away from it. Once we are aware of the threat and have time to consider our actions we swat it. (Or shoo it out of the door in my mum's case)

 

[ducati916]

pedrochemical

It seems that you are looking into ergodic/dynamical systems, and statistical mechanics.

Without going technical, probability theory & Benoit Mandelbrot, with his fractal mathematics constitute the body of work that have been applied to stockmarkets.

Most experienced chartists will pick the stockmarket data over random generations. There are a number of tests available that you can take. Do I buy into charts? Yes, I do, but not in the way that technical analysts approach the problem.

As you progress, there will be one major problem to overcome. If you can solve it, then, you'll make some serious money...if not, well there you go. Timeframes and psychology are central to the problem, and to the solution, therefore no argument from me.

jog on
duc

 

[pedrochemical]

What I fear is that I have again come across something deterministic but non-computable.
This universe sucks sometimes!!

 

[Trembling Hand]

What I fear is that I have again come across something deterministic but non-computable.
This universe sucks sometimes!!

If you are looking for a game that will enable you to be proven right, for whatever reason, there is no more dangerous game than the markets.

 

[motorway]

pedrochemical

It seems that you are looking into ergodic/dynamical systems, and statistical mechanics.

Without going technical, probability theory & Benoit Mandelbrot, with his fractal mathematics constitute the body of work that have been applied to stockmarkets.

Most experienced chartists will pick the stockmarket data over random generations. There are a number of tests available that you can take. Do I buy into charts? Yes, I do, but not in the way that technical analysts approach the problem.

As you progress, there will be one major problem to overcome. If you can solve it, then, you'll make some serious money...if not, well there you go. Timeframes and psychology are central to the problem, and to the solution, therefore no argument from me.

jog on
duc

And What does mandelbrot say about how to treat time and time frames ?


On Ergodic

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=636484

Living in a non-ergodic world

once ergodicity is broken, we
are guaranteed at least two different estimates for the returns, that is,
one given by temporal considerations (i.e., a time average), and another
generated through stuctural assesments (i.e., an ensemble average). In
general, we can think that in a non-ergodic market, every subset of
publicly available information produces a different estimated return,

In other words, in
non-ergodic markets some information sets will generate a better estimate of future returns than others.

motorway

 

[pedrochemical]

If you are looking for a game that will enable you to be proven right, for whatever reason, there is no more dangerous game than the markets.

I do not want to be proved right or wrong - I just want to discover some stuff and save some of my more gullible friends from seeing stuff and believing in stuff that isn't actually there.
I would never presume to be able to apply what I am talking about to actual trading - I am not that arrogant or well funded or clever and I only trade with my money- it also keeps the mind ticking over while I 'slowly bleed to death' waiting until the next great disaster day to fill up on.
I am sure we can all agree - in terms of making cash from the markets, 2008 was the best year in living memory. So why, then, have I a lot of friends who lost money in the largest year on record. I want to know (and I want them to know) why they did lose when they followed conventional wisdom.
I have my suspicions and I am working at discovering if I am right.

Anyway, at the moment I think it makes sense to stay liquid.

 

[Trembling Hand]

I am sure we can all agree - in terms of making cash from the markets, 2008 was the best year in living memory. So why, then, have I a lot of friends who lost money in the largest year on record. I want to know (and I want them to know) why they did lose when they followed conventional wisdom.
I have my suspicions and I am working at discovering if I am right.

Yes agree but could you explain more on the underlined please?

Anyway, at the moment I think it makes sense to stay liquid.

Doesn't it always

 

[pedrochemical]

By conventional wisdom, I mean they traded trends as they would when they were making money. Then they started losing money - on stronger trends.
That seems crazy to me.

I worry that the reason they were making money in 2005-2007 was not because their chartist convictions were correct - this is what I want to confirm somehow - hence my pestering of the board for opinions.

I do not think it was entirely arrogance or gullibility either. I think maybe the notion that we can predict the future was fundamentally flawed.
It is the maybe that I want to turn into a yes or no or at least a probably or probably not

In some cases I can imagine people believe the general trend of a market is up as this is what they see day in day out for years. Perhaps they are naturally more reluctant to exploit the bear market - perhaps this a human trait?
But a lot of non-stupid traders stopped making money too - so that ain't it.

I do not even beleive in certain cases (where they were used) their 'bots were incorrectly configured.
They were not caught out by a tidal wave - I think that maybe their education told them that what happened was so unlikely that they could not see it when it was happening. Who knows?

But I am getting off the point here.
I want to know if we can predict the future - and get it right.
And if we can/can't - then why?

This is not a zero sum game so I am unhappy when people lose money unnecessarily.

 

[motorway]

I want to know if we can predict the future - and get it right.

FEEDFORWARD

The fact you make a prediction , and (at least someone ) acts on it
( important )

That changes the Future

Multiply this by all the people predicting something and acting on it.
And you have to input all these into your model

and then input that into your model and so on like a hall of mirrors to infinity.

NOW that was why all the people blew up..

BUT if can predict something and very few can act on it ( or would act on it )
The you have an opportunity..

You loan actions will not at first be disruptive..

Such opportunities arise eg in a bear mkt

so many know the market will go up again
But they have NO MONEY
or worse are being sold out

The guy that lost the Kingdom for want of a horse
He knew where and what he should be doing

But constraints meant the difference

The difference between feedforward and feedback..

From a distance I can see what shopping cue is the shortest
But acting on that information means that when I arrive it has become the longest cue..

We are the thing we are trying to predict

Not like the behaviour of rabbits or microbes or gasses ,coin tossing or roulette wheels

motorway

 

[Cartman]

By conventional wisdom, I mean they traded trends as they would when they were making money. Then they started losing money - on stronger trends.
That seems crazy to me.

interesting thread

Pedro ---- the reason why your mates started losing is due to them not recognising the trend was changing, and adapting to it -----

it was their perception which was flawed --- they were seeing what they wanted to see; not what was actually happening, hence holding onto bad positions for too long --- their positions ended up being a bi-product of basic fear and greed --- it just happens unfortunately

I want to know if we can predict the future - and get it right.
And if we can/can't - then why?

This is not a zero sum game so I am unhappy when people lose money unnecessarily.

depends -- how far into the future do you want to be able to predict??

and really how beneficial would it be to know that XYZ will be $X in december 2010 if in the meantime its gona have a trading range of 2000 pips/ticks/whatever ----
if you want to use strong binoculars, your probably gona need deep pockets

i dont like seeing people lose either, but it is the nature of the game --- if your mates are smart they should learn a lot from their losses and hopefully come back stronger next time. Cheers.

 

[ducati916]

motorway

Ergodic theory is a branch of mathematics that studies dynamical systems with an invariant measure. In mathematics, an invariant measure is a measure that is preserved by some function. The mathematical concept of a function expresses dependence between two quantities, one of which is known and the other which is produced.

Ergodic systems, statistically speaking, [the system] that evolves for a long time "forgets" its initial state.

Clearly, financial markets do not conform to this theory. Our function is the two variables, time [known] and price [unknown/future] which qualifies the financial markets, as time and price are dependent. However, Markets, tend, to mean revert thus invalidating a definition of an ergodic system.

I looked at your link, but fail to ascertain the point you are making.

pedro

By conventional wisdom, I mean they traded trends as they would when they were making money. Then they started losing money - on stronger trends.
That seems crazy to me.

I worry that the reason they were making money in 2005-2007 was not because their chartist convictions were correct - this is what I want to confirm somehow - hence my pestering of the board for opinions.

If you look at the S&P500 chart, 2005-2007 was a period of the lowest volatility on record. Buy & Hold, was the correct strategy for that time period. Contrast the current period, the volatility increased to all time record highs for sustained volatility.

While the trend was sustained [in hindsight] the ability to hold became very difficult. Thus, making money became much harder.

jog on
duc

 

[motorway]

Clearly, financial markets do not conform to this theory.

That was the point .

I misunderstood you

I thought you were implying the opposite.

motorway

 

[tech/a]

Is there a difference to the chartist...to a good one, who is cognizent of the difference, yes. To the poorly informed, none whatsoever.

This is more what I was getting at.

Now don't laugh, but I have a lot of data and am in the process of taking a Poincare Section.
My computer is not fast enough, however, to draw the pretty pictures.
The progression through phase space (vector field style) of what I am seeing is what I was expecting to see. This makes me nervous as an experimenter, but even though the numbers are as yet muddy, I hope to see the difference between randomly generated sequences and historic sequences.

I shall let y'all know what happens.
Pretty old school maths but applied like I am trying it - could be interesting if not novel.

Thanks for the input.

My son is a Doctor of Physics.
He hasnt the time to do any modelling on the markets currently.
However as a long term Chartist myself he has watched and taken part in some of my trading.

When asked how he feels about mathamatical modelling with regard to trading his reply was un expected.

"I can return you figures in any format you want---to support an arguement or not."

 

[Trembling Hand]

"I can return you figures in any format you want---to support an arguement or not."

Exactly. And to a trader that is of absolutely no use. We will just and move onto making money from the random moves

 

[motorway]

So where does the shopping cue problem lead to..

The more others are acting on the same information to predict
and inform ...

Even if they are using complicated indicators
moving average. oscillators. bands . Gann EW etc

The patterns will break down just at the point where they start working well
( you end up at the end of the day in the long cue )


They break down Into dampening or expanding congestion patterns

With Cobweb ( from original Hog Cycle pattern ) dynamics .



What is the solution ???

One is to be quick , very quick .

The other is to focus on the key characteristic that most/all the mathematical models overlook..

One person walking in towards the cues
Knows something everybody does not

He has been told that a cue over the far side will be opening in the next few minutes.. And he positions himself to take advantage..

In the markets positioning oneself means buying or selling FIRST

ie: Accumulation or Distribution



motorway