Fat Tails, Power Laws, & Bumble Bees

[ducati998]

Humans are a social species, and therefore we are all connected to each other through a variety of ways.

This is especially true in regards to financial matters. Traders & Investors respond not only to relatively objective economic considerations, but also to political and pure speculation or rumour.

Movements in stock prices on a daily basis are generally limited to 1% to 3%. This would describe the Gaussian curve for price movement.
However, stock prices will also move 10%, 20%, and sometimes even more. These price movements are supposedly statistically rare.

However, Mandlebrot, clarified that these so called anomalies, were actually a lot more common than statistically suggested. Hence the phenomena of fat-tails

What has been subsequently hypothesised, is the reason for this fat-tail phenomenon. And this is the power law

The power law relates to connectivity and networks. Informal network studies, initially, or partly came about from the Erdos law
Erdos, was a prominant mathematician, who wrote possibly hundreds of papers, often collaborating, or having original work updated.
This led to a system of detailing how close to the original work, the subsequent mathematician was, and assigning an "Erdos" number.

Power laws were adapted by Barabasi, to measure the size of the internet. Anyway, the relevance to the market, fat-tails, and price movements is this.

If the Barabasi model is correct, then the financial networks have a definite shape and inter-connectedness. Most of the time, this makes no difference, and "price" stays within the Gaussian curve, or normal distribution.

However, in times of high volume, the trades become very strongly influenced by a relatively few central nodes to the network, which causes a reinforcement of the initial move.

High volume is a proxy for herd thinking, with the inter-communication and reinforcement of the same idea.

Of course this is a corollory to "when all think the same, no-one thinks at all".
However, a monitoring of the nodes, may, or possibly provide insight as to future outcomes, if of course you place any faith in this sort of thing in the first place.

cheers d998

 

[chump]

For all of the above read ..momentum and quants analysis ....

 

[ducati998]

chump

Yes, in part, however what I wish to do is differentiate from the slugs & ants thread as to how, who and why, these distortions are driven.

cheers d998

 

[chump]

Hmm.well I should have posted my last post on here then

but you still have the wrong end of the stick when it comes to other methodologies..let me try to make my point using an example from ordinary business practice..

For the sake of this example...imagine you are a retailer ..you sell bars of chocolate...you sell 100,000 a year ...on any given day you get 500 feet through your door buying bars of chocolate...

now you notice that the 500 feet through the door has dropped off..last month the average was only 350 ...now if you are FA based you will wait to your period end whenever that may be..and you will find you sold approx 30% fewer bars of chocolate than you expected..

however , if you took the observation of feet through the door to be TA observations you would have found out much earlier that your expectations were going to be confounded...and this is the equivalent of TA observations because feet through the door is no part of your fundamental data as you would find it in a set of accounts..it is an indirect event that will form an outcome that you will eventually see in your accounts.. just as the observation of a lack of momentum on a chart will eventually be seen to be reflected in the future outcome to price on that chart...

Look the best I can do to express this is we are not trading accounts ..we are selling bars of chocolate for profit and we want to know if we are not selling as many as expected..likewise for a trade ..we are not trading accounts we are buying or selling price for profit..and we want to know when that price expectation is not going to be delivered.

what your approach is doing is no more and no less than ignoring and washing away the noise that occurs within your timeframe for a trade..in other words it's keeping you in the trade longer than might be the case if you implemented a different trade management (stop approach)...but you could have done the same using either TA or Quants .it's about observation of what you are trading and the forming of a strategy for that instrument that leads to profit ..nothing more ..the methodology is irrelevant ..

My footnote on this is you are guilty of confusing poor application with poor methodology...

 

[Tuffty]

The power law is explained in the book 'A mathematician plays the market' by John Allen Paulos.

Spooky, The power law really seems to have some odd effects. The first post on this thread is very similar in structure and words to the book form page 175 on!! Must be one of those fat tails!

 

[ducati998]

Tuffty

Nothing spooky at all, that is exactly where I pinched the idea from. He introduces the idea, but leaves it somewhat undeveloped.

There are some interesting areas that have a real relevance to trading, or investing.

cheers d998

 

[ducati998]

chump

You are certainly giving it a fair crack of the whip, however, where your analogy is a little weak is in the fact that you are using a corner shop type of business.

When I say this, it is not the product at this point that is the issue, but rather the financial structure of the company. Corporations, have very different dynamics due to their capitalisation structure amongst many others.

All businesses will suffer swings through business and economic cycles, this is standard. The financial condition of the business is one of the important factors in determining how that storm is weathered.

however , if you took the observation of feet through the door to be TA observations you would have found out much earlier that your expectations were going to be confounded...and this is the equivalent of TA observations because feet through the door is no part of your fundamental data as you would find it in a set of accounts..it is an indirect event that will form an outcome that you will eventually see in your accounts.. just as the observation of a lack of momentum on a chart will eventually be seen to be reflected in the future outcome to price on that chart...

However you have touched on the central tenet of this thread, and that is via "networks" or other factors, yet to be discussed, how is new information released, how is it spun, who gets to act upon it first..........those with the first look, how are they analysing this information, how influential are they in others decisions.............and, do they form the herds opinion.

If they do, then the above questions may interest you. Regarding TA "leading" from FA, does it?
I personally disagree. I feel TA is so lagging, as to be trying to run through treacle.

what your approach is doing is no more and no less than ignoring and washing away the noise that occurs within your timeframe for a trade..in other words it's keeping you in the trade longer than might be the case if you implemented a different trade management (stop approach)...but you could have done the same using either TA or Quants .it's about observation of what you are trading and the forming of a strategy for that instrument that leads to profit ..nothing more ..the methodology is irrelevant ..

Here I strongly disagree. The methodology that can sweep, or remove noise, is by a very wide margin the more profitable.

Noise will cost you a great deal of money in failed trades. The entire point of a stoploss, is however to prevent potentially even larger losses.

Notice the emphasis on losses.
This constant attrition of capital will grind down all but the superior traders, as it tends to be frequent. The tighter the stop, the greater the number of times it will be hit, especially in higher beta securities.

Technical traders cannot remove the stoploss in traditional TA, as of course TA measures sentiment, nothing else.

In the other thread, we discussed methodologies that did not rely on measuring sentiment, and the consistent commonality, was the ablation of the stoploss concept as a money management technique ( I accept your modified concept, but this is not an option for a technical trader )

My footnote on this is you are guilty of confusing poor application with poor methodology...

Actually, I just posted the complete opposite on the other thread......................spooky.
Cheers d998

 

[chump]

LOl..this could run and run...

"The methodology that can sweep, or remove noise, is by a very wide margin the more profitable."...and here is where you keep getting the wrong slant ...it is not the methodology that does this ..that is, it is not whether you adopt a FA or TA approach that determines this ...it is the strategy within your methodology...do you hear the rush of gold coins coming down the chute ? ..it is the strategy ...take FA ..adopt the wrong strategic approach in the way you use it and you're going nowhere ..same with TA....

hence we come back to the point I have already made ..it is not the methodology ..it is the poor application that can not come through with a winning strategy and on the other thread you have come close to saying the same thing when you talk about poor practice ...

My analogy is not weak at all and business large or small (and I've been in both) is always about getting the numbers right at the end ..how you do that might vary ,but for the topic in hand the analogy was adequate...

 

[ducati998]

chump

it is not the methodology that does this ..that is, it is not whether you adopt a FA or TA approach that determines this ...it is the strategy within your methodology...do you hear the rush of gold coins coming down the chute ? ..it is the strategy ...take FA ..adopt the wrong strategic approach in the way you use it and you're going nowhere ..same with TA....

Ok, I see where you are coming from now...............( I think )
Certainly, if your chosen strategy within your chosen methodology is faulty, you will have problems.

As by way of illustration, methodology, fundamental, strategy, growth stocks if your skill in identifying growth stocks is suspect, or weak, or just plain wrong, you will struggle. Agreed.

However, and this is where we still differ, my assertion is that, Technical analysis is so flawed as a methodology, it becomes very difficult ( not impossible ) to implement, or even define a successful strategy.

So, the point of my threads, was to highlight an area that had possibly been overlooked, viz market efficiencies, or inefficiencies, as a place to consider looking at for improving a strategy within the Technical methodology.

My analogy is not weak at all and business large or small (and I've been in both) is always about getting the numbers right at the end ..how you do that might vary ,but for the topic in hand the analogy was adequate...

Which is my point. The numbers can be manipulated in many legitimate ways by GAAP, some positively, some negatively, that distiguishing them in reality can give a huge advantage. A bad number, can be a massive opportunity. Whereas, in a small business, a bad number is usually just that ............a bad number.

Cheers d998

 

[ducati998]

In examining the effect, ( if any ) we can take a recent, popular, and highly successful IPO.
GOOG

Now GOOG, as far as I know was if not the first, one of the very few businesses to break with Wall St tradition, and cut out the investment banks of their slice of the action.

They were to say the least, not best pleased. In fact there followed a lengthy campaign by all the investment banks, to discredit GOOG and the process of their flotation, valuation, and business model.

One can only speculate as to their position had they secured the business.
GOOG was floated, at $85 per share, massively down on initial expectations of $140 per share

Round one to Wall St.
Having succeeded in destroying some 40% of the initial valuation, here comes the hype, as of course, at $85, and the initial destruction of value to the owners, Wall St could jump in an "undervalued" popular IPO, and move this.

Who's collusion would be necessary?
Well Funds, hedge, or mutuals. Pension Funds, in fact the whole gamut of the inside network.
Sure enough, the shares are now $300 odd.

Where are the nodes of control?
Who may signal that the party is over, or just beginning?
How can you spot it..............is it even possible, or by definition must you be inside?
If unspottable, then how can you pick up the early signs, that it is game over?

Currently we are in Round 2
Has it reached boiling point, or still heating up?

cheers d998